MalignantTumors ./ turn//~ ,/

KENNETH ENG ENCLOSE WITH ITEM Return Date

23.FEB~ 4

3 G3

/See User's Handbook CU 5 4 6 3 4

Y LOAN

ONLY

Return to: The British Library Lend=ng D~v~ston,Boston Spa, Wetherby, West Yorkshire LS23 7BQ if no other library indicated,

TABLE SURGICAL ANATO,M Y

.

Sphincters

.

OF

.

.

.

.

.

.

.

.

.

.

.

Nerves

.

.

.

.

.

.

.

.

Blood Supply . .

CONTENTS .

. .

.

. .

.

.

PilYSIOLOGY OF ANAL CONTINENCE

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

6

.

.

.

8

.

.

.

. . .

. .

.

.

.

.

.

.

6

.

.

9

.

.

Sphincter Continence

.

.

.

.

.

.

.

.

.

.

.

.

Reservoir Continence

.

.

.

.

.

.

.

.

.

.

.

.

LYMPIIATICS

.

.

.

.

.

CANCER OF TIlE RECTUM

.

.

.

Etiology and Epidemiology Polyps and Cancer Adenomatous Carcinoma

.

.

.

Lymphoma

.

Leiomyosarcoma

.

. .

.

. .

.

.

.

.

Resection

Anterior Resection

.

.

.

Abdomino-Anal

Pull:Through

Abdorninosacral

Resection

SURGICAL MANAGEMENT Choice of Operation

.

.

.

.

16

.

.

.

17

.

.

.

.

.

.

.

18

.

.

.

.

19

.

.

.

.

.

.

.

.

20

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

. .

.

.

.

.

.

.

.

. .

. . . .

23 23

.

24 24

.

.

21 22

. .

Resection .

.

.

Resection .

.

. .

.

. . .

.

.

22

.

.

.

.

.

. .

. . .

.

Rehabilitation After Abdominoperineal SUMMARY

15

.

Resection

Technique of Abdominoperineal

15

.

.

.

Technique of Anterior Resection Technique of Abdominosacral

.

.

.

.

.

.

14

.

.

.

.

.

.

.

.

.

14

.

.

14

.

Resection

.

.

.

.

.

. .

.

.

.

.

.

.

. .

. .

.

.

12 .

.

.

.

.

.

EVALUATION OF OPERATIVE APPROACIt Abdominoperineal

.

.

10 11

. .

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

. .

.

.

.

. .

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

. .

.

.

.

.

. .

.

.

.

.

. .

.

.

. .

Melanoma

.

.

.

.

. .

.

.

.

.

.

Polyps

Villous Tumor Carcinoid .

.

.

l0

.

. .

.

.

.

.

. .

25

.

29 36 41

.

42

QUESTIONS OF CLINICAL USEFULNESS ANSWERED IN THIS ISSUE 1. What is the estimated incidence of carcinoma of the colon in the United States? p. 5 2. What is the most common site? p. 5 3. What are the anatomy and function of the puborectalis muscle and the external sphincter? pp. 6 - 8 4. How is the physiology of sphincter and reservoir continence related to function following the various types of rectal resection? pp. 1 0 - 1 2

5. What is the significance of the lymph node studies of McKay, Westhues, Gabriel, Dukes, Gilchrist and David, Grinnell, Best, Blair and others? pp. 1 2 - 1 4 6. What are the principal known predisposing factors in the etiology and epidemiology of cancer of the rectum? p. 14 7. What is the incidence of malignancy in adenomatous polyps? p. 15

8. What is the incidence of malignancy in villous tumors? pp. 15 and 16

9. On the basis of studies of distal intramural axial spread of cancer, what is considered a safe surgical margin? p. 16 10. What are the indications and limitations of electrocoaguiation, radiation and chemotherapy as primary therapy for carcinoma of the rectum? p. 17 11. What are the indications and limitations of electrocoagulation, radiation and chemotherapy as adjunctive therapy for carcinoma of the rectum? p. 17

12. Describe the course, therapy and prognosis of carcinoids, primary malignant lymphoma, leiomyosarcoma and malignant melanoma of the rectum, pp. 17-21 13. What were the significant contributions of the early surgeons to the present procedures for surgical treatment of malignant tumors of the rectum? pp. 2 2 - 2 3 14. What are the indications for anterior resection? pp. 14 and 24 15. What are the indications for abdominosacral resection? pp. 15 and 24

16. What are the indications for abdominoperineal resection? pp. 16 and 24

17. What are the important surgical principles in prevention of ureteral injury and anastomotie leak in rectal resections? pp. 2 5 - 3 6 18. How is a colostomy matured, a paracolostomy hernia prevented and rapid perineal healing ensured? pp. 37--40 19. What are the essentials of rehabilitation of the patient following abdominoperineal resection? pp. 41 and 42

4

is Professor of Surgery at New York University School of Medicine, Attending Surgeon at University Hospital and Visiting Surgeon at Bellevue Hospital, New York. Doctor Localio received his M.D. degree from the University of Rochester and has been the Johnson and Johnson Distinguished Professor of Surgery since 1972. Surgery of the alimentary tract is his principal clinical and research interest.

is Assistant Professor of Surgery at the New York University School of Medicine. Doctor Eng received his M.D. degree from that institution in 1967. He serves as Attending Surgeon at University Hospital and Director of the 'A' Service at Bellevue Hospital, New York.

T H E L A R G E I N T E S T I N E is the most common site for r0alignancy of the gastrointestinal tract and ranks second only to the lung in incidence of cancer deaths. The American Cancer Society estimates that 99,000 new cases and 49,000 deaths will be attributed to this disease in the United States in 1975.z About half of these tumors will be found in the rectum, an area that is readily accessible to diagnosis by physical examination or proctoscopy but that presents some special surgical considerations. Proper therapy for malignancies of the rectum includes early diagnosis and complete extirpation of the lesion, with preservation of sphincter function when appropriate. When sacrifice of the sphincters is essential for cure or effective palliation, it is incumbent upon the surgeon to provide rehabilitation of these individuals. Successful management of patients with malignant lesions of the rectum is dependent upon a detailed knowledge of anatomy, an understanding of anal continence and an appreciation of the mechanism of spread of malignancy. NOTE: This work was supported in part by a grant from The Billy Rose Foundation, New York, New York.

SURGICAL ANATOMY The rcctosigmoid junction is variable but generally lies at the.level of the third sacral vertebra. The rectum conforms to the sacrococcygeal curve and is bordered anteriorly by the urinary bladder, seminal vesicles and prostate, or by the female reproductive organs (Fig. 1). The rectum pierces a diaphragm of striated muscle across the pelvis to become the anal canal and ends in an external opening buttressed by the external and internal sphincters and lined by squamous epithelium. This apparatus of exquisite function can differentiate and sclectively permit the passage of flatus while forming a barrier to the passage of liquids and solids. SPItINCTERS Courtney,2~ Goligher e t a l p 2 and Oh and Kark 82 have reviewed and restudied the anatomy of the anal sphincters. The internal sphincter is Fig. 1.--Relation of the rectum to pelvic organs and sphincter apparatus.

~P TMENT

~ICtAL r~.~ENT

Fig. 2.--Anal sphincters. The internal sphincter is a thickening of the circular smooth muscle of the rectum. The external sphincter consists of a superficial compartment formed by the subcutaneous and superficial external sphincters and a deep compartment formed by the intermingled fibers of the deep external sphincter and puborectalis.

Fig. 3.--Pelvic diaphragm showing the puborectalis sling. ',TALIS "OR .~.NI

7

a thickening of the circular smooth muscle layer of the rectum (Fig. 2). This is surrounded by the external sphincter, consisting of a superficial compartment formed by the subcutaneous and superficial external sphincter muscles and a deep compartment formed by the intermingled fibers of the deep external sphincter and the puborectalis. The puborectalis a.rises from the posterior aspect of the pubis and passes posteriorly around the anorectal junction to join with the puborectalis from the other side to form a sling (Fig. 3). At the upper border of the anal canal the confluence of the internal and external sphincters and the puborectalis sling forms the anorectal ring, a structure that must be preserved to ensure continence following any operative procedure in the anal region. BLOOD SUPPLY The rectum receives blood from the superior hemorrhoidal artery and the paired middle hemorrhoidal and inferior hemorrhoidal arteries (Fig. 4). The superior hemorrhoidal artery arises as a direct continuation of the inferior mesenteric artery in the base of the sigmold mesentery. At the rectosigmoid junction it divides into two branches, which turn laterally and descend on either side of the rectum to supply the rectum and anal canal and anastomose with the middle and inferior hemorrhoidal arteries. The middle hemorrhoidal arteries are branches of the internal iliacs, Fig. 4.--Blood supply and lymphatic drainage of the rectum. Arrow indicates the primary zone of lymphatic spread of rectal cancer--upward along the superior hemorrhoidal vessels.

il)t , LYMPFtATICS

IA N

~ ~ , ~ , ~ :i

i

E "

_ Im

"

"

9 SUPER OR HEMORRHODAL "

MIDDLE HEMORRHOIDAL ARTERY

9 INFERIOR HEMORRHOIDAL~ NODES

BLOOD SUPPLY

R

I

R

2

D

F

~INFER[OR

H I ALARTERY HEMORR O O "

which run medially from the pelvic wall below the pelvic peritoneum in the lateral ligaments of the rectum. The inferior hemorrhoidal arteries are derived from the internal pudendal branches of the internal iliac arteries. They run medially in the ischiorectal fossae to supply the sphincters and anal mucosa. The extensive anastomoses between the inferior, middle and superior hemorrhoidal vessels provide adequate blood supply to the rectal stump from the inferior hemorrhoidal artery alone, when the middle and superior hemorrhoidal vessels are sacrificed. In mobilization of the proximal colon for low anastomoses, adequate blood supply is preserved following high ligation of the superior hemorrhoidal and left colic arteries or the inferior mesenteric artery at its source, provided the marginal artery is preserved and there is no pre-existing arteriosclerotic occlusion. NERVES The rectum is innervated by both sympathetic and parasympathetic components of the autonomic nervous system. Sympathetic preganglionic fibers arising from the first three lumbar segments pass through the sympathetic chains to synapse in the preaortic plexus and through the hypogastric nerves to synapse in the right and left presacral or hypogastric plexuses. Postganglionic fibers from the preaortic plexus follow the course of the inferior mesenteric artery to the upper rectum, whereas fibers from the hypogastric plexus pass directly to the lower rectum, anal canal and other pelvic viscera. Injury to the sympathetic nerves may occur whenever the presacral space is entered and will result in loss of the ability to ejaculate. Parasympathetic innervation is derived from the nervi erigentes that emerge from the second, third and fourth sacral foramina and pass through the presacral plexus and upward through the preaortic plexus to synapse in the myenteric and submucosal plexuses. These nerves are liable to injury when the lower rectum is mobilized and the lateral ligaments of the rectum divided. Such injuries will result in neurogenic bladder or impotence, or both. The internal sphincter, like the rectum, is innervated by autonomic nerves. The external sphincter and levator ani are voluntary muscles supplied by branches of the pudendal nerve from the fourth sacral nerve. Somatic sensory innervation of the anal canal arises from the same source and reaches 0.5-1.5 cm craniad to the anal valves. 37 The internal sphincter relaxes with rectal distention, an intramural reflex that is maintained following spinal paralysis or regional anesthesia to the perirectal plexus. The external sphincter and puborectalis contract in response to rectal distention, a spinal reflex that is accompanied by cortical awareness of pressure and the urge to defecate. These important reflexes will be considered further in the discussion of anal continence.

PHYSIOLOGY OF ANAL CONTINENCE Duthie 3s states that anal continence depends upon an adaptable barrier formed at the anorectal junction and in the anal canal by a combination of forces due in part to the configuration of the region and in part to the action of the muscles. These forces are activated in response to sensory information from the rectum and anal canal. Certain sensory information accompanies the passage of material from the sigmoid to the rectum. The cortical representation of the information is dependent upon conditioning and social habits. Reflex relaxation of the upper anal canal permits contact of colon content with the sensitive anal mucosa. The abundantly innervated epithelium signals the type of content and the urge to evacuate. Continence is the ability to subvert this urge to the will. Reflex contraction of the external sphincter and the puborectalis sling elevates the anal canal, elongates it and pulls it forward, increasing the angle between the rectum and anal canal converting the orifice to an anterior-posterior slit. Additional support may be added by voluntary contraction of these muscles, which may be maintained for 1 or 2 minutes or until the urge to defecate passes.

SPHINCTER CONTINENCE Current concepts of the anatomy of the sphincter apparatus and the physiology of continence emphasize the preservation of (1) the sensitive mucosa of the anal canal, (2) receptors in the lower third of the rectum, (3) innervation of the anorectal region and (4) the sphincter apparatus including the critical puborcctalis. Operations designed to preserve sphincter function that do not preserve these structures will be associated with less than satisfactory function'---hence (1) the failure of pull-through operations that preserve the muscle but remove the sensitive anorectal mucosa--and (2) the uneven results of those operations that disrupt innervation of the anal canal by extensive dissection and eversion of the anorectal stump for anastomosis. Duthie and Gairns 37 have demonstrated a profuse innervation in the wall of the anal canal, especially in the region of the anal valves. Free epithelial and organized nerve endings mediating exquisite sensitivity to pain, temperature, touch and pressure give an anatomic basis for preservation of the anal canal and its epithelium. The exact nature and the need for preservation of receptors in the lower rectum are less precisely established. Based on balloon studies in patients following abdomino-anal resection, Goligher and Hughes 5~ suggest that a 6-8 cm rectal stump is required for preservation of the rectal warning mechanism and the rectosphincteric reflex. However, in patients with Hirschsprung's disease, Schtirli and Kiesewetter9z have demonstrated preservation of sensation and the rectosphincterie reflex when the colon is brought down through the puborectalis sling to the anal canal lo

even when a portion of the internal sphincter is sacrificed. They place the sensory receptors in the puborectalis or in the perirectal tissue at this level. However, their observations may be dependent on the greater adaptability of a pediatric population. In our own experience abdominosacral resection with preservation of a 2-3 cm rectal stump is sufficient to preserve the warning mechanism and anal continence.65 This corresponds to resection at the level of the levators, with preservation of the puborectalis and the sphincters. The anastomosis is performed posteriorly above the pelvic floor, obviating the need for extensive mobilization, stretching and eversion of the anal canal. These observations may reconcile the difference in the findings of Goligher and Kiesewetter. Goligher's finding that a 6--8 cm rectal stump is necessary for continence is based upon the assumption that the receptors lie in the rectal wall and are removed with low resections. An alternative explanation is that the receptors are extrarectal and their innervation is disrupted by mobilization of the rcctum and eversion of the anal canal for the exterior anastomosis. In abdominal pull-through operations for Hirschsprung's disease in which mobilization is carefully performed close to the rectal wall and in abdominosacral resections .in which the anal canal, its surrounding musculature and nerve supply are not disturbed, continence is maintained following resection of the lower rectum and anastomosis at the level of the levators. RESERVOIR CONTINENCE The concepts of reservoir continence and sphincter continence were ably delineated by GastonJ 4 To fully evaluate the results of sphinctersaving operations, it is important to understand and differentiate these concepts. The left colon is a reservoir organ that, like other smooth muscle conduits, relaxes, to accommodate intraluminal content. Relaxation of the smooth muscle allows the colon to accumulate bulk to a certainmagnitude; pressure then rises and peristalsis and propulsion of the fecal mass occur. If there is no sphincter to breach, evacuation occurs. This is the course of events in the patient with a colostomy and underlies our method for rehabilitation of the colostomy patient. Not infrequently in such individuals appreciable periods of time may elapse between episodes of colonic emptying, and, indeed, the patient may develop a rhythmic periodicity that can be augmented by food or by the instillation of fluid into the stoma. This reservoir continence demonstrable in an abdominal colostomy may be achieved in a perineal colostomy and the resultant effect erroneously attributed to retained sphincters. In some reported series reservoir function has undoubtedly been mistaken for sphincter continence. It must be emphasized that such reservoir function is not under voluntary cortical control and can be breached by involuntary colonic motility. 11

Removal of the sphincters, destruction of the puborectalis, denervation of the anorectal region or removal of the anal mucosa will be associated with sphincter incontinence. Preservation of these structures and preservation of an ample portion of rectum, as in anterior resection, will be associated with both sphincteric and reservoir continence. Very low resections bringing proximal colon to the lower rectum or anorectal junction have in our experience been associated with sphincteric continence but with aberrations in reservoir function. 65 The sensitive sensory epithelium of the anal canal is exposed to frequent stimulation, and repeated small movements are apt to occur during the waking hours until some reservoir function returns. It has been our observation that reservoir function returns more rapidly if the stool contains bulk and moisture and if the patient can repress the minor stimuli for the need to evacuate.

LYMPHATICS The rectum, like the remainder of the colon, has two intercommunicating lymphatic pathways. The intramural lymphatics consist of lymphatic vessels and lymphoid follicles in the submucosal and subserosal tissues that drain into the extramural lymphatics. The extramural lymphatics consist of lymph vessels and lymph nodes that tend to follow the veins and arteries and provide anatomic boundaries for surgical resection (see Fig. 4). Lymph vessels may be interrupted by nodal structures in the areolar tissue surrounding the longitudinal muscles, the pararectal nodes of Gerota. The latter lie above the levator muscle in close proximity to the posterolateral wall of the extraperitoneal rectum. Efferents also pass directly to the intermediate nodes and principal nodes along the superior hemorrhoidal and inferior mesenteric arteries. This upward route of drainage with its nodal barriers is the principal zone of spread for rectal tumors and lends itself to surgical ablation. Dye studies have shown that lymphatic drainage from points above the middle rectal valve, 7-8 cm above the anal verge, is exclusively upward. 1~ Below the 7-8 cm level, lymph may drain superiorly or laterally along the lateral rectal ligaments to the middle hemorrhoidal nodes or directly to the parietal lymphatics of the lateral pelvic wall and the iliac vessels. It must be emphasized that these lateral lymphatics lie in the areolar tissue beneath the pelvic peritoneum and bear no relationship to the levator ani muscles. The anal canal is drained by lymphatics that run with the inferior hemorrhoidal vessels in the ischiorcctal space as well as with the middle hemorrhoidal and superior hemorrhoidal pathways. Other lymphatics cross the genital femoral sulcus to drain into the superficial inguinal nodes. The reader is referred to the chapter on lymphatics of the large intestine by Slanetz and Herter 9~ for further details. Confronted with the high recurrence rate following perineal and 12

sacral operations for rectal cancer. Miles 76 developed the abdominoperineal resection. His observation on the upward zone of spread of cancer was sound. It has become apparent that his views of the lateral and inferior zones of spread that required sacrifice of sphincters and wide resection of the levators were based largely on the study of patients with extensive carcinomatosis with obstruction of the proximal lymphatics and nodes and retrograde lateral and inferior spread. The Miles concept continues to dominate surgical practice, despite the fact that his observations, in his own words, were based on " . . . post-mortem findings in regards to spread of cancer In patients who died from advanced and inoperable cancer. ''77 In 1920, McVay r5 a fellow at the Mayo Foundation, studied 100 cases of lymph node involvement in cancer of the rectum: 47 patients had nodal involvement, and only one patient had involvement below the lower limits of the lesion--a single node at a distance of 1 cm. This observation was confirmed by Westhues 1~ in 1930; Gabriel, Dukes and Bussey a3 in 1935; Gilchrist and David 4~ in 1938; Coller, Kay and Maclntyre zl in 1940; Grinnell ~5 in 1942 and Glover and Waugh 46 in 1946. Best and Blair 11analyzed the combined data of the above authors. In 607 combined cases, 12 ( 2 . 3 % ) had nodal involvement 1-2 cm below the lesions and five patients ( 0 . 8 % ) had nodes 2 cm below the lesion. The case for lateral spread along the levators is dependent on the position of the lesion and the extent of upward spread. Coller, Kay and Maclntyre zl failed to find metastases in the lateral zone of spread in lesions that were 3 cm or more above the anorectal junction. Guernsey, Waugh and Dockerty 57 studied 133 cases of rectal cancer that were 6 cm or less above the insertion of the superior surface of the levator: 42 were 2 cm or less above the insertion and in no case was there evidence of spread along the levators or the lateral lymphatic chain. Dafoe z7 studied 100 specimens of advanced rectal cancer located within 4.5 cm of the anorectal junction and found only one instance of lateral spread along the levator ani. As early as 1935, Gabriel, Dukes and Bussey, 43 on the basis of careful dissection of nodes in 100 specimens and examination of over 2,000 glands, concluded that " . . . lateral or downward lymphatic spread is only found in a later stage of the disease when hemorrhoidal lymphatics are choked by metastases." Correlation of the lymphatic anatomy of the pelvis, experimental studies on the lymphatic drainage at different levels in the rectum and examination of pathologic specimens in patients with rectal Cancer provide the basis for evaluating the adequacy of operations for rectal cancer. These studies demonstrate that the drainage of midrectal cancer is primarily upward and that the lateral route of spread is rare even in the face of far advanced disease. Furthermore, the middle hemorrhoidal nodes lie in the lateral ligaments of the rectum and the pelvirectal areolar tissue, bear no relation to the levator ani muscle 13

and are encompassed by wide mobilization of the rectum above the levator ani muscles. Lymphatic spread beyond these confines is lateral to the parietal lymphatics of the pelvic wall and lilac vessels and is beyond the scope of all commonly practiccd radical resections.: All the evidence indicates that, except in far advanced cases that are beyond surgical cure, cancer of the midrectum may be brought under control without sacrificing the levator ani and the sphincters. This is supported by the long-term data available for sphincter-saving operations.V, ,2, ,4. io~ Lesions below the midrectum, at the anorectal junction and at the anal canal require abdominoperineal resection. For some anal tumors, inguinal lymph node dissection will be necessary to encompass the regional lymphatic spread.

CANCER OF THE RECTUM

ETIOLOGY AND EPIDEMIOLOGY Although the exact cause of cancer of the rectum is unknown, certain predisposing factors and premalignant conditions have been identified. Epidemiologic studies show a predilection for western populations, the incidence of cancer of the large bowel being greater in North America and Europe than in Asia and Africa, whereas the incidence in Japanese Americans and American Negroes approaches that of the white population of the United States. Aries e l al. r' made the intriguing observation that the feces of English subjects contain significantly more bacterioids and fewer streptococci and lactobacilli than those of Ugandan Africans, indicated that bacterioids are extremely active in the degradation of bile salts and suggested this as a possible source of carcinogens. Burkitt ~v emphasized the lack of fiber in the diet of western man, which results in slow transit and a greater potential for action.of fecal carcinogens on colorectal mucosa. The association of adenomatous polyps with cancer and the premalignant nature of villous tumors, familial polyposis and ulcerative colitis are well established, s0

POLYPS AND CANCER The controversy concerning the malignant potential of adenomatous polyps has continued for almost two decades. 94 Castleman and Krichstein, '9 Ackerman and Spratt ' and Wheat and Ackerman 10s are of the opinion that adenomatous polyps are distinct from villous tumors and have negligible malignant potential. Dukes, 3~ Ewing, 39 Morson TM and Sunderland and Binkley ~00 consider adenomatous polyps and villous tumors to be two differing growth patterns of" the same premalignant lesion. Whereas most adenomatous polyps are firm, spherical, lobulated, pedunculated tumors and most villous tumors are 14

soft, shaggy, sessile lesions consisting of numerous finger-like processes, cytologic characteristics are tile same. Some large sessile polyps have the configuration of adenomas, and some villous tumors are pedunculated. Transitional morphology and mixed histology are common. Likewise, all stages of malignant degeneration from focal atypia and carcinoma in situ to invasive cancer have been identified. Finally, 10% of resected cancers will show contiguous areas of benign tumor. 78 The question of management of polypoid lesions has been the subject of numerous statistical studies. Review of 1856 polypoid lesions by Grinnell and Lane ~Gshowed an over-all incidence of invasive caneer of 6 % . The incidence of cancer in adenomatous polyps was 2.9% and in villous tumors 31.9%. Less than I % of adenomatous polyps 1 cm or less in diameter and at least 9 % of those 1.5 cm or larger showed invasive cancer. Metastasis to regional lymph nodes did not occur until the cancer had penetrated the base of the stalk beyond the muscularis mucosa. ADENOMATOUS POLYPS Although the origin of carcinoma from polyps is controversial, there is little disagreement concerning the management of polypoid lesions. Ackerman and Spratt t observed that, since it is impossible to distinguish radiographically or by inspection between adenomatous polyps and polypoid carcinoma, all lesions should be removed. With the advent of fiberoptic colonoscopy, all pedunculated polyps have come within the reach of the electrocautery snare. Bowel rescction is reserved for lesions with histologically proved invasion of the submucosa, or anaplastic carcinoma. Small sessile adenomatous polyps lying below the peritoneal reflection may be removed via sigmoidoscope. Higher sessile lesions generally require segmental resection, with the exception that those less than 1 cm in diameter can probably be safely biopsied and observed, in view of the low incidence of malignancy in polyps of this size. VILLOUS TUMOR Whatever the biologic relationship between adenomatous polyps and villous tumor the high incidence of malignancy in the latter has been well documented. Grinnell and Lane 5G identified invasivc cancer in 31.9% of 216 villous tumors. Wheat and Ackerman l~ found focal atypia, focal carcinoma or carcinoma in 74% of 50 specimens. Clearly, the villous lesion has great potential for malignancy and we suggest that it be designated villous tumor. The term villous adenoma might connote a lesser danger similar to that of adenomatous p o l y p . . Discovery of a villous tumor of the rectum should alert the surgeon to the possibility of a second lesion, frequently a frank carcinoma higher in the colon. 15

The focal nature of malignancy in the large villous tumor makes sigmoidoscopy, even with multiple biopsies, unreliable in resolving the question of malignancy. An area of ulceration, induration or fixation suggests malignancy and the necessity for radical resection even in the face of negative biopsies. For clinically benign lesions, total excisional biopsy is the minimal acceptable therapy. Small lesions in the extraperitoneal rectum may be amenable to local excision, whereas those lying above the peritoneal reflection will require segmental resection. Larger lesions or suspicious lesions of the midrectum or higher are ideallysuited for removal by abdominosacral or anterior resection. Abdominoperineal resection may be necessary for large, circumferential, low-lying lesions when there is great likelihood of local recurrence, postoperative anal stricture or incontinence. CARCINOMA As alluded to in the discussion of lymphatic spread, many investigators have amplified and refined the original observation of Miles in the pathology of rectal cancer. Dukes 3~ demonstrated a correlation between local extent of tumor and the incidence of lymphatic and venous spread. His system of classification, based upon 1,000 rcsccted specimens, has been widely accepted as a guide to prognosis and a means of evaluating surgical results. Class A lesions, those in which the tumor is confined to the rectal wall with no nodal metastases, are found in 15% of specimens, and virtually all are curable by resection. Class B cases, with tumor invasion of perircctal tissues but negative nodes, are observed in 35% and are associated with 70% successful treatment. Class C lesions, those with positive nodes, constitute 50% of cases, and no more than 30% of these will be cured of disease. Dukes 3G found a close correlation between histologic grading and staging by extent of invasion. Tumors of high grade malignancy such as seen frequently in young patients are highly invasive and associated with early dissemination. In contrast to lesions of the stomach and esophagus, carcinoma of the rectum seldom spreads far beyond the grossly apparent margins. Dukes 36 states, "The margin of the tumour is usually so distinct and the colour contrast so sharp that the first impression of the extent of the growth formed during the dissection has not often to be modified by the subsequent verdict of the microscope." Black and Waugh 15 studied longitudinal spread in 103 specimens and found microscopic extension of 0.5 cm or more in four cases and 1.2 cm in only one case. Therefore it is likely that a distal surgical margin of 3 cm is sufficient for control of intramural axial spread, although wider margins are taken when possible. The mainstay of therapy for carcinoma of the rectum is surgical resection. Other modalities that have been applied as adjunctive or 16

primary therapy are radiation, chemotherapy and electrocoagulation. Papillon s4 advocates definitive treatment of rectal cancer by endocavitary radiation but emphasizes that this is strictly a local method for early cancers without lymphatic spread. It is difficult to accept the contention that the presence of nodal metastases can be accurately assessed by digital examination. Stearns e t al. 95 reported in 1959 a retrospective study that showed a tendency for improved survival with preoperative irradiation in patients with Dukes C, but not A or B rectal cancer. Subsequent prospective studies by the same authors 96 failed to confirm the value of this therapy. Other randomized prospective studies of preoperative irradiation are now in progress. The use of preoperative irradiation in 15 patients to render unresectable tumors operable has been reported by Kligerman and UrdanettaLaffee.6~ We have had experience in three such cases. One tumor was judged unresectable at laparotomy and was subsequently removed with great difficulty following irradiation. The other two were irradiated because they were judged clinically to be inoperable. The method was successful in one patient, resulting in obvious decrease in size and making subsequent abdominoperineal resection possible. The value of radiation therapy appears to be limited to palliation of inoperable lesions or of recurrences in the pelvis and bone. Chemotherapy for gastrointestinal cancer using 5-fluorouracil, an inhibitor of DNA synthesis, has been widely used for the treatment of recurrences and as an adjuvant to operation. Response rates are variable, and to date there is no prospective randomized study demonstrating improved survival. Electrocoagulation of rectal cancer was first used by Strauss e t al. 9s,99 in 1913 initially for palliation but subsequently as the primary treatment. More recently, the method has been used by Madden and Kandalaft71,72 and Crile and TurnbuIl} 2 There are no valid longterm data comparing this procedure with abdominoperineal resection, nor is such information forthcotriing, since electrocoagulation, unless accompanied by laparotomy to ascertain nodal or hepatic involvement, precludes pathologic staging. The suggestion that radical resection is unnecessary is based upon Grinnell and Lane's ~6 finding that high ligation of the inferior mesentcrie artery held no advantage over ligation at the superior hemorrhoidal pedicle, and ignores the welldocumented advances of Lockhart-Mummery's 6s wide perineal excision and Miles ';6 abdominoperineal resection. At present we would consider electrocoagulation only for the patient who presents an unacceptable operative risk.

CARCINOID In 1888, Lubarsch ;0 described an ileal tumor arising from the crypts of Lieberkiihn, and his report still stands as the classic description of the carcinoid. The true incidence of rectal carcinoid is unknown but 17

some 869 cases have been reported since Saltykow 91 first discovered a carcinoid at this site in 1912. Caldarola e t al. TM and Quan e t al. 8s report an incidence of approximately 0.04% on routine sigmoidoscopy. Carcinoids of the rectum are generally asymptomatic and most often present as small, hard, gray or yellow submucosal nodules found incidentally on sigmoidoscopic examination. Malignant carcinoids usually are larger and may ulcerate or assume a polypoid and rarely an annular configuration, in which case rectal bleeding, change in bowel habit or varying degrees of obstruction may occur. Unlike ileal carcinoids, which are often multiple, the rectal lesions usually are solitary. The affinity for silver stains, which led Gosset and Masson 53 to identify these tumors as arising from Kultchitsky's cells, is rarely seen in rectal carcinoids. The carcinoid syndrome that may result from massive hepatic metastases has to our knowledge been documented only once in association with rectal carcinoidP 0 The clinical behavior of rectal carcinoids is more closely correlated with size and gross appearance than with histologic appearance. Empirically, small carcinoids 1 cm or less in diameter may be locally excised, as evidenced by Caldarola's TM series of 129 such lesions treated by excision or fulguration with no recurrences 1 to 13 years later. Carcinoids that reach 2 cm or more in diameter often are grossly malignant and frequently exhibit deep invasion and lymphatic or hematogenous metastases at the time of diagnosis. In a series of 25 rectal carcinoids reported by Peskin and Orloff, 86 nine of 10 patients with lesions 2-2.9 cm in diameter exhibited local invasion or lymphatic metastases. Six patients bad distant metastases when first seen and died with widespread disease. Of the three patients treated by abdominoperineal resection, two were alive and. free of disease. Large rectal carcinoids and those with grossly malignant appearance should be treated by abdominoperineal or radical sphincter-saving resection. Lesions less than 2 "cm may be treated initially by local excision and subjected to careful pathologic examination. Invasion into the muscularis propria is an acceptable indication for radical operation. Treatment must be individualized, with the understanding that no clear distinction exists between benign and malignant lesions. LYMPHOMA Malignant tumors of lymphoid tissue may occur in the rectum as manifestations of generalized malignant lymphoma, as primary tumors or as malignant lymphoid polyposis affecting long segments of the alimentary canal, including the upper gastrointestinal tract. Primary malignant lymphoma of the rectum is exceedingly rare. Lymphomas arise in the submucosa and may protrude into the lumen or invade the gut wall. Thus, polypoid, plaque-like or annular lesions, with or without ulceration, may be seen. Stout 97 points out 18

that, unlike carcinoma, these lesions seldom produce a desmoplastic response and retain some flexibility. All histologic types including lymphosarcoma, reticulum cell sarcoma and Hodgkin's have been described. Symptoms, when present, are similar to those produced by carcinoma, namely, bleeding, change in bowel habit or obstruction. Analysis of prognosis and treatment of choice for primary malignant lymphoma of the rectum are complicated by the rarity of the lesion and lack of uniform criteria for diagnosis in the literature. Helwig and Hansen ~8 point out that benign growths probably have been incorrectly diagnosed as malignant, and Dawson e t al. 29 suggest that many lymphomas thought to be primary in the rectum are simply a local manifestation of generalized disease. Although these tumors are radiosensitive, recurrence is the rule after irradiation alone. In the Mayo Clinic series of 10 malignant lymphomas of the rectum, there were no cures in the seven patients treated by irradiation alone and only two patients survived longer than 9 months. The only long-term survivor was alive and well 7 years following local excision and irradiation.24 Surgical excision appears to offer the best prospect for cure. Stout97 reported one long-term survivor alive and well 13 years after abdominoperineal rcsection in a group of 37 patients with colorectal lymphosarcoma treated by various modalities. Dawson e t al. 29 reviewed 37 cases of primary malignant intestinal lymphoma. Of the 12 patients with lesions localized to the rectum, two were alive 15 years or longer following surgical excision. Grossly enlarged lymph nodes were often free of tumor, and regional lymph node involvement bore little correlation with prognosis. Recently Perry e t al. 85 reviewed their experience with 22 patients with primary lymphoma of the rectum. Eight of 10 patients who underwent chemotherapy, radiation or no ,treatment died, with poor survival times. Five of nine patients were alive and well following abdominoperineal or anterior resection. Three patients were irradiated following preliminary colostomy. Two of the three responded favorably and subsequently had colostomy closure. The available data clearly favor radical surgical excision with or without irradiation for primary malignant lymphoma of the rectum. Unresectable lesions should be treated by preliminary colostomy and irradiation. Chemotherapy is the logical choice for patients in whom the rectal lesion is a manifestation of generalized lymphoma. LEIOMYOSARCOMA These rare tumors arise from the smooth muscle of the muscularis propria, muscularis mucosa or blood vessels of the rectum. The lesion is submucosal but, as it enlarges, may project into the lumen or outward to impinge upon and invade adjacent structures. Unlike smooth muscle tumors of the stomach, lesions in the rectum are as apt to be malignant as benign. Leiomyosarcoma is highly invasive locally and 19

metastasizes by way of the blood stream, rarely via regional lymphatics. Histologic examination reveals a more or less orderly arrangement of smooth muscle cells. Bizarre cells are not a reliable indicator of malignancy nor is evidence of local invasion, since these smooth muscle tumors arc not encapsulated. The prevalence of mitotic activity probably is the best sign of malignancy. Symptoms, when present, are due to pressure from the enlarging mass and ulceration of the overlying mucosa. Cystic degeneration may lead to the incorrect diagnosis of perirectal abscess. Leiomyosarcoma is resistant to irradiation and must be treated by radical surgical excision. Although there have been reports of successful treatment by local excision, recurrence is the rule following this procedure. Eight of 10 patients at the Mayo Clinic treated by local excision with or without radiation therapy had recurrence. 4 The experience reported by Morson 79 from St. Marks hospital was essentially the same. The treatment of choice is radical operation by abdominoperineal or sphincter-saving resection, with emphasis on wide excision of perirectal tissue in view of the great propensity for local invasion. The ultimate prognosis is poor. Although 25% of patients will survive for 5 years, subsequent deaths from distant metastases are common. MELANOMA Although there have been reports of primary malignant melanoma of the rectum, most authors 'G. Go,81. s3, s9 agree that these tumors arise from the squamous epithelium of the anal canal. Since preferential growth is upward, the tumors usually present in the lower third of the rectum. If the growth is protuberant, the tumors may be easily confused with thrombosed herriorrhoids. Nonpigmented forms are common, making differentiation from simple polyps, skin tags or fissure difficult. Larger tumors may be hard to distinguish from other malignancies at this site. Histology is variable but the presence of pigment or tumor giant cells is helpful in making the diagnosis. The common symptoms of rectal bleeding, diarrhea and tenesmus are nonspecific. High index of suspicion followed by biopsy is essential to diagnosis of this rare tumor. Malignant melanoma is highly invasive and metastasizes early via the blood stream and via regional lymphatics to the superior hemorrhoidal nodes, the lateral wall of the pelvis and the inguinal nodes. Surgical removal offers the only chance of cure, but prognosis is still poor. Braastad e t al. 1G reviewed the literature in 1949 and found only three patients (each treated by local excision) followed long enough to be considered cured. Quan e t al. 89 reported no 5-year sur20

vivors among 21 patients, although those treated by radical resection enjoyed a longer survival time. Pack and Martins 83 favored abdominoperineal resection with simultaneous bilateral groin dissection, but could report only one survivor alive and well 10 years later in a group of eight patients treated in this manner. Husa and Hockerstedt6~ reviewed the subject recently and concluded that prognosis is related to stage of disease rather than to type of treatment. They report one long-term survivor (treated by abdominoperineal resection) from a series of 14 patients treated by various methods. Abdominoperineal resection followed by bilateral groin dissection should still be used in patients with melanoma without evidence of distant metastases, although the value of this approach is by no means established.

EVOLUTION OF OPERATIVE APPROACH The earliest approaches to carcinoma of the rectum were the perineal excisions reported by Faget in 1739 ~~and by Lisfranc in 1826. 63 Since general anesthesia was not yet available, one must recognize the trepidation of the surgeon and the fortitude of the subject. It is not surprising that subsequent efforts were not reported for 50 years until VerneuiP 0z described his method of perineal excision, including excision of the coccyx for wider exposure. In England the Allinghams2 suggested an inguinal colostomy for control of the inevitable perineal sepsis and practiced this procedure until the early 1900s. Lockhart-Mummery69 made significant improvements in the adequacy of this operation and in the quality of survival. Like the Ailinghams, he used preliminary colostomy but extended the dissection intrapcritoneally to include division of the levators and mobilization between rectum and sacrum posteriorly and the prostate, seminal vesicles and bladder anteriorly. The lateral ligaments were divided and bowel was pulled down to divide the superior hemorrhoidal vessels and rectum as high as possible. This operation, although inadequate for the upward zone of spread, was practiced for 20 years at St. Marks Hospital in London with a mortality of 11.6% and a 5-year survival of 40% .42 On the continent, surgery for rectal cancer took a somewhat different course. In 1885, Kraske 62 reported to the 14th Congress of the German Association of Surgeons his approach to rectal cancer by excision of the coccyx and lower sacrum for posterior exposure of the rectum. The operation consisted of segmental resection with end-toend anastomosis or excision of the entire rectum, and sacral colostomy. Hochenegg59 explored different methods for constructing a more secure anastomosis. One method consisted of eversion of the rectal stump and a pull-through of the proximal bowel for external anastomosis. The completed anastomosis was replaced above the sphincter. Hochenegg's durchzug method involved removal of the anal mucosa and a pull-through of proximal bowel with the hope that the serosa 21

would adhere to the raw surface of the anorectal stump. These sacral operations, inadequate for the upward zone of spread, associated with a large, slowly healing defect and a tenuous anastomosis or an inaccessible colostomy, were widely practiced. Mahdi 73 reviewed the experience from Hochenegg's clinic and reported a mortality of 12% and a 5-year survival of 30%. ABDOMINOPERINEAL RESECTION Removal of the rectum by combined abdominal and perineal operation was first performed in 1883 by Czerny,26 who entered the abdomen to complete the local excision of a lesion not removable through a sacral incision. Others had practiced the combined procedure, but it was Miles 76,77 who first defined a radical operation to encompass the lymphatic spread along the pelvic mesocolon in addition to wide periheal excision. Based upon his personal experience of 54 recurrences following 59 perineal excisions and a careful study of the spread of rectal cancer, he advocated abdominoperineal resection in every case of carcinoma of the rectum. The original efforts were associated with a higher mortality (40%) and morbidity than the Lockhart-Mummery procedure. Miles's abdominoperineal resection was slow in acceptance by the surgical community until the importance of the upward zone of spread, lower mortality and better 5-year survival were established. The Miles operation with minor technical modifications has become the standard operative procedure for carcinoma of the lower rectum. MaunselF 4 first suggested, and Weir 1~ practiced, an abdominal resection of the rectum with preservation of the anorectal stump. The stump was evaginated, proximal colon was brought through and an exterior anastomosis established in a manner similar to that described by Hochenegg. Studies on the lymphatic spread of cancer by McVay, 75 Westhues, 107 and Gabriel e t al. 43 and others gave new impetus to sphincter-saving operations: (1) anterior resection in which resection of the upper rectum and anastomosis are achieved entirely through the abdomen, (2) pull-through procedures in which anastomosis is achieved at the perineum and (3) abdominosacral resection involving posterior anastomosis. ANTERIOR RESECTION As far as can be determined, the first anterior resection was performed in England by Cripps in 189728 and in this country by Balfour in 1910.10 These operations and that of Lockhart-Mummery69 and others were entirely intra-abdominal, with restoration of continuity by telescopic tube technique. The present day anterior resection with sutured end-to-end anastomosis was extensively practiced by Dixon 3~ and the group at the Mayo Clinic. Their data and those of others 22

clearly demonstrated that its effectiveness was comparable to that of abdominoperineal resection for cure of carcinoma of the upper rectum. ABDOMINO-ANAL PULL-TtlROUGtI RESECTION The Maunsell-Weir type of pull-through resection with anastomosis to an everted anorectal stump was critically reviewed by Goligher, 4s who found that dependable rectal function returned in most patients in 4 - 6 months. He suggested proximal transverse colostomy in all cases in view of the high incidence of anastomotic breakdown. 49 Among the modifications aimed at avoiding anastomotic leakage were those devised by Babcoekfi Black and Botham 14 and Bacon. s Babcock's and Black and Botham's operations entailed drawing the proximal colon through the anus without suture anastomosis. Excess colon was excised later, after spontaneous union of the cut edge of the anorectal stump and serosa of the colon. However, failure of union and necrosis of the protruding gut presented serious problems. Babcock practiced section of the anal sphincter to protect the blood supply to the colon, a maneuver that impaired anal continence. Bacon restored continuity in a manner sinfilar to Hochenegg's d u r c h z u g method, drawing colon through denuded anorectal stump to provide a wider area for union. Sacrifice of the anal mucosal receptors, however, results in defective anal continence. A more recent a b d o m i n a l pull-through procedure described by Turnbull and Cuthbertson TM and Cutait and Figlioni 25 is essentially a two-stage Maunsell-Weir operation in which the proximal colon is pulled through an everted anal stump, but excess colon is not amputated until secure healing of the anastomosis has occurred. Goligher et aL so reported a series of 40 patients, with a mortality'of 12.5% and a major necrosis of the colon stump in 22.4%. Although 68% of the patients were satisfied with the over-all result, 75% had defective control of flatus, 61% had occasional fecal soiling and 71% wore a protective pad. The large series of Bacon s and Waugh TM have established the adequacy for cancer resection of various pull:through procedures. Functional results in terms of anal continence have been less consistent. ABDOMINOSACRAL RESECTION Abdominosacral resection is based upon the sacral excision of the rectum described by Kraske 62 in 1885. Years later, recognizing the frequency of abdonfinal recurrences following sacral excision, he favored upward extension of the resection by an abdominal approach. This operation with various modifications has been practiced by Finsterer, 4~ Goetze, 47 d'Allaines and de Vernejour, 2s Donaldson et al. 3~ and Localio. 6~, 65 The abdominal phase of the operation includes high 23

ligation of the superior hemorrhoidal vessels and wide mobilization of the sigmoid colon and rectum, as in the other sphincter-saving operations. The bowel is drawn through the posterior wound, resection and anastomosis achieved above the pelvic floor without disturbing the sphincters, puborectalis sling, anal mucosa and innervation of these structures. As in all low rectal anastomoses, this operation has been complicated by troublesome fecal fistula, and a temporary colostomy is recommended by Donaldson. 3a In our experience carcinoma of the midrectum may be treated by abdominosacral resection without routine colostomy with acceptable morbidity, normal postoperative continence and long-term survival equal to abdominoperineal or anterior resection. ~5 SURGICAL MANAGEMENT CttOICE OF OPERATION Anterior resection with primary anastomosis is the accepted procedure for malignancy of the upper third of the rectum. Cancer of the lower two thirds is still generally treated by abdominoperineal resection. Studies on lymphatic spread and many large clinical series have established the rationale and acceptable survival rates for sphincter preservation in treatment of midrectal cancer compared to conventional abdominoperineal resection. The transsacral approach is a logical means for restoration of intestinal continuity after radical resection of midrectal cancer. In our experience 6~-66 the direct posterior exposure of the distal limit of resection above the pelvic floor allows accurate construction of the anastomosis without disturbing the anorectal structures and their innervation, has been attended by essentially normal anal continence and is preferable to the various pull-through techni~lues. The position of the lesion from the anal verge measured by two observers at different times has been the principal determinant for choice of operation. Lesions of the midrectum, defined by the lower limit of the tumor falling within 5.5-10 cm of the anal verge in women and 7-1 1 cm in men, are considered suitable for abdominosacral resection. Lesions above these limits are treated b y anterior resection and primary anastomosis, and lesions below the midrectal level are treated by conventional abdominoperineal resection (Fig. 5). Fixed cancer is considered inoperable by any of these procedures. Hepatic metastases have not been a contraindication to resection. Histologie grading, although related to prognosis, has not altered our approach, and this plan has been followed'for all rcmovable growths. Measurement of the level of the lesion by sigmoidoscopy has certain limitations. A loose extraperineal rectum with S-shaped curves may yield measurements considerably shorter than actually exist "after mobilization and straightening of the bowel. Measurement in the kneechest position will decrease errors. 2,1

AN"([ ~IOR RESECTION MEN;ABOVE t I cm WOMEN;ABOVE 10cm

AI]DOMINOSACRAL MEN:7-11 r WOMEN. $,5-10 r

ABDOMINOPERINEAL MEN: BELOW 7 cm WOMEN: BELOW 5 5 Cm

Fig. 5.--Extent of resection. A, anterior resection includes superior and middle hemorrhoidal vessels and lymphatics. B, abdominosacral resection extends dissection to the level of the levator anl. C, abdominoperineal reseclion removes, in addition, levator ani, anorectal stump and inferior hemorrhoidal I~'mphatics. Shaded areas indicate extent of resection. Darker shading indicates zone of tumor.

TECHNIQUE OF ANTERIOR RESECTION A left paramedian incision is made from above the umbilicus to the pubis (Fig. 6). After exploration the small bowel is walled off and kept out of the field by the Trendelenburg position. The sigmoid loop is drawn into the wound and freed by dividing filmy adhesions to the lateral wall of the pelvis. The peritoneum at the base of the mesosigmoid is incised on the left, and the gonadal vessels and the left ureter are identified as they cross the lilac vessels. Dissection in the loose areolar plane in front of the iliac vessels and aorta lifts the sigmoid mesocolon. A similar peritoneal incision is made on the right, and the right ureter is identified and swept laterally. The superior hemorrhoidal vessels are then divided and ligated at the level of the aortic bifurcation. The presacral space is now entered and, with upward traction on 25

Fig. 6.--Left paramedian incision for abdominoperineal or anterior resection. Dashed line indicates extension for mobilization of splenic flexure.

the sigmoid colon, the lower sigmoid and the rectum are freed from the anterior surface of the sacrum by blunt dissection down to the tip of the coccyx (Fig. 7). Several dense connective tissue bands may require division with scissors. Care should be taken to stay in the proper areolar plane to avoid bleeding from torn presacral veins. The peritoneal incisions are continued anteriorly to meet in the rectovesical or rectovaginal pouch. The areolar plane on the anterior surface of the rectum is entered and dissection carried down to the prostate or to the posterior wall of the vagina (Fig. 8). With anterior and posterior attachments separated, the lateral ligaFig. 7.--Anterior resection. The rectum has been freed from the sacrum and drawn caudad. Aortic bifurcation and ureters are clearly seen.

RECTU~

SACRU~J

AORTA

26

Fig. 8.--Anterlor resection. Anterior dissection showing separation of rectum from bladder and seminal vesicles.

ments of the rectum containing the middle hemorrhoidal vessels are easily visualized, divided and ligated. Careful hemostasis is a~hieved at this stage. A right angle clamp is placed on the rectum 4 cm below the lower limit of the lesion, and the rectum is divided. The proximal margin for resection near the junction between the descending colon and sigmoid is selected, the bowel divided and the specimen removed. An open two-layer anastomosis is then performed using interrupted nonabsorbable 4-0 seromuscular suture and a continuous 4-0 chromic catgut suture for the inner layer (Fig. 9). The basic surgical principles of constructing a watertight anastomosis under no tension, with preservation of an adequate blood supply, have been widely preached. We believe that this is best accomplished by an open two-layer inverting anastomosis using fine suture material. To place this anastomosis in a rigid pelvic cavity bathed in blood and serum is to invite disaster. Our practice is to take utmost care that hemostasis is achieved and that mobilization of the proximal colon is sufficient to allow the bowel to conform loosely to the sacral curve. This usually requires mobilization of the splenic flexure and often 27

O

\'-

Fig. 9.--Anterior resection. A, division of the rectum. Note the right angle clamp applied at least 4 cm below the lowest palpable margin of the lesion. B, open two-layer anastomosis showing insertion of posterior row of interrupted nonabsorbable sutures. C, completed anastomosis covered by omentum.

28

requires division of the left colic artery. The pelvic floor is not reconstructed. The omentum is brought down to and around the anastomosis, and loops of small bowel are allowed to fill the pelvic cavity, obliterating dead space and sealing the anastomosis. With these precautions we have found that drainage of the pelvis is unnecessary. Anastomotic dehiscence following low anterior resection may be as common as 69% when leaks are actively sought by digital examination, sigmoidoscopy or barium enema. 50 Clearly, major anastomotic leaks resulting in clinically significant signs and requiring diversion of the fecal stream are far less frequent. Too early and overzealous examination of the suture line probably is unwise. During the past 6 years, only two (1.4%) anastomotic leaks have been detected in 139 consecutive anterior resections. 65 Although it is well established that carcinoma of the rectum seldom spreads beyond the grossly apparent margins, suture line recurrence by implantation of exfoliated tumor cells is a real danger. Our practice has been to irrigate thoroughly with normal saline after careful isolation of the two open ends of the bowel prior to construction of the anastomosis. TECHNIQUE OF ABDOMINOSACRALRESECTION The patient is placed in the lateral position with the left side up. This position affords good abdominal exposure for mobilization of the left colon and splenic flexure and for pelvic dissection. If the mobilized lesion is sufficiently high to permit anterior resection and anastomosis, this can be and has been readily accomplished. For abdominosacral resection, simultaneous posterior exposure with the abdomen open facilitates delivery of the proximal colon and allows further mobilization if necessary for a secure anastomosis w~thout tension. Finally, if abdominoperineal resection is necessary, the perineal dissection may be completed without repositioning the patient. The incision is started to the right of the midline above the pubis, carried transversely acr.oss the rectus and then obliquely, parallel to 9

P

Fig. lO.--Abdominosacral resection. The patient is placed in the left lateral position. An oblique abdominal incision is indicated prior to draping. The abdominal phase of the operation is performed from the right side of table.

i

29

Fig. 1 1 .--Abdominosacral resection. The sigmoid colon is drawn downward to the patient's right. The peritoneum at the base of the mesosigmoid is incised exposing the left ureter.

the inguinal ligament, to end at the midaxillary line between the tip of the twelfth rib and the iliac crest (Fig. 10). The incision is deepened through all layers and the abdomen is explored. 'The small bowel is drawn out of the peritoneal cavity and appropriately protected. The peritoneum of. the left lumbar gutter is incised down to the brim of the pelvis and the left ureter identified (Fig. 11 ). The omentum is separated from the left transverse colon at the embryonic fusion plane. The splenic flexure is now readily mobilized (Fig. 12). The mesosigmoid is lifted, the peritoneum on the right incised and the right ureter identified (Fig. 13). The superior hemorrhoidal artery is ligated at its source along with the accompanying vein (Fig. 14). If the colon and mesocolon are short, the left colic artery is ligated at its source, with care taken to protect the marginal vessels and preserve the blood supply from the middle colic artery. The presacral space is entered and the rectum is mobilized down to the levators posteriorly (see Fig. 7). The peritoneal incisions are carried into the cul-de-sac and the anterior surface of the rectum is freed to the prostate or the middle third of the vagina (see Fig. 8). Traction on the bowel now exposes the lateral ligaments of the rectum for division. The left is more easily visualized and is divided first, and the 30

SPLEEN

f ~

Fig. 12.--Abdominosacral resection. The splenic flexure is well visualized and readily mobilized at the lateral aspect of the oblique abdominal incision.

Fig. 13.--Abdomlnosacral resection. Identification of the right ureter. Assistant holds the slgmoid colon upward to the patient's left.

R]GHT URETER

31

SUPERIOR HEMORRHOJDALARTERY LEFT COLrC ART INFERIOR ME,~ENTERIC ARTERY

Fig. 14.--Abdominosacral resection. The superior hemorrhoidal vessels are divided. The left colic artery, shown intact, may be sacrificed for greater mobilization of proximal colon.

Fig. 15.--Abdominosacral resection. MobilizaHon of the rectum has been completed, as shown in Figs. 7 and 8. Division of the left lateral ligament of the recturn showing middle hemorrhoidal arlery.

32

Fig. 16.--Abdomlnosacral resection. Division of the colon at Ihe proximal limit of resection between CopeDeMartel clamps.

middle hemorrhoidal vessels identified and ligated (Fig. 15). The right lateral ligament is now more easily divided and secured. Hemostasis is essential and is accomplished at this stage. The colon is transected at the junction of the dcscending colon and sigmoid using the Cope-DeMartel clamp (Fig. 16). A transverse incision is made at the level of the sacrococcygeal joint. The tip of the coccyx is pressed forward to render its attachment to the sacrum prominent, and the joint is divided with a scalpel. The fascial and muscular attachments of the coccyx are divided and the bone is removed, giving access to the presacral space (Fig. 17). The wound is enlarged by splitting the posterior levators and, if necessary, the medial portions of the gluteus maximus may be incised. The distal stump is delivered, and additional dissection is performed to prepare the bowel for transection and anastomosis at least 3 cm below the lesion (Fig. 18). A right angle clamp is applied. The proximal colon is delivered, with care taken that proper orientation is maintained. We do not hesitate to mobilize the proximal colon' further if there is any question of tension on the anastomosis. The anastomosis is constructed in two layers using interrupted 4-0, nonabsorbable sutures for the outer layer and continuous 4-0 chromic catgut for the inner layer. The anterior row of sutures is placed and tied before division of the bowel. The anastomosis is completed as depicted in Figure 19. The absence of tension, a good blood supply, fin6 suture material and an adequate lumen are essential to protect against leakage. To this we add, when possible, the protection of the omentum. The pelvis and posterior wound are irrigated with saline. Hemostasis is secured, the pelvic peritoneum is left widely open and the wounds are closed without drainage. When colostomy is deemed necessary, the transverse colon is divided and separate stomas brought out through the rectus muscle. In the first 50 abdominosacral resections, seven ( 1 4 % ) anastomotic leaks were discovered. 66 The one early death in this group was due to fecal peritonitis. Six of seven leaks occurred in men 42-56 33

Q

t

~,CYGEAL JOINT

Fig. 17.--Abdominosacral resection. A, the position of the sacrococcygeal joint is determined by forward pressure on the coccyx and marked prior to draping. B, transverse incision at the level of the sacrococcygeal joint. C, depression of the coccyx makes the sacrococcygeal joint prominent, and disarticulat~on with a knife is easily accomplished in the right plane. D, removal of the coccyx.

years of age. We believe that a leak is related to technical difficulties encountered in the muscled young man with a small pelvis. We now recommend diverting colostomy in men below the age of 65 years and in other cases when the integrity of the anastomosis is in doubt. When the postoperative course is uncomplicated, the colostomy is closed during the same hospitalization. We continue to perform one-stage operations in women and in older men. 34

Fig. 18.--Abdominosacral resection. The specimen has been delivered through the posterior wound. A right angle clamp is applied 3 cm below the I~sion.

35

0

0

RECTUI

\

6)

./4

Fig. 19.--Abdominosacral resection. Anastomosis. A, proximal bowel has been delivered, care taken to maintain proper orientation. Anterior row of interrupted nonabsorbable sutures is inserted. B, division of the bowel. C, anterior continuous row is started in the middle, using a doublearmed chromic suture and run in both directions. D, completion of continuous catgut inner layer. E, completed anastomosis covered by omentum.

ii OMENTUMCOVERSCOMPLETED ANASTOMOSIS

36

TECHNIQUE OF ABDOMINOPERINEAL RESECTION

Our technique for abdominoperineal resection is essentially the same as described by Miles with certain refinements that we believe facilitate postoperative recovery and improve ultimate rehabilitation. 67 The abdominal phase of the operation is performed through a lower left paramedian incision. Mobilization of the left colon and rectum is accomplished essentially as practiced by Miles and is described in detail in the technique for anterior resection. The mesosigmoid is incised, vessels ligated and the sigmoid colon divided using the CopeDeMartel clamp. The clamped distal sigmoid is tucked into the presacral space. The pelvic peritoneum is neither mobilized nor closed, Fig. 20.--Abdominoperineal resection. Formation of the colostomy. A, the colon is sutured to all layers of the abdominal wall. B, the wound has been closed. The mesocolon is divided. C, excess colon is trimmed 1-2 cm above skin level. D, the stoma is sutured to the skin.

O

O

|

37

permitting the omentum to be drawn down over the small bowel to fill the pelvis. The proximal end of the colon is brought through the upper end of the incision and carefully sutured to the edges of the peritonea ! and fascial layers with 4-0 chromic catgut sutures placed in the mesosigmold, appendices epiploicae, teniae coli and sometimes the seromuscular layer itself. These sutures should not penetrate mucosa, to protect against the possibility of fistula. The abdomen is closed using buried monofilament sutures. Protruding colon is trimmed to 1-2 cm above skin level and then sutured to the skin edge using interrupted 4-0 chromic catgut taking full thickness of colon and dermal layer of the skin (Fig. 20). The patient is turned to a right lateral Sims's position, the upper buttock is retracted with adhesive tape and a pursestring suture is placed to close the anus (Fig. 21). An elliptical incision is made around the anus from the central point of the perineum to the tip of the coccyx. As the incision is deepened through the subcutaneous tissue and fat of the ischiorectal fossae, branches of the inferior hemorrhoidal vessels are ligated. The anococcygeal raphe is inciscd at the tip of the coccyx, and by finger dissection the presacral space is entered. This opening is enlarged by incising the levators on cither side, and the sigmoid and rectum are delivered (Fig. 22). Incisions in the levator ani are carried forward, dividing the muscles between clamps. In the anterior plane of dissection developed during the abdominal phase, the specimen is easily separated from the prostate and membranous urethra or posterior wall of the vagina. The puborectalis and rectourethralis muscles are divided and the specimen Fig. 21.--Abdominoperineal resection. Position and incision for the perlneal phase. The anus is closed with pursestring suture.

38

\ Fig. 22.-~Abdomlnoperineal resection. A, the incision has been deepened to the levators. Anococcygeal raphe is incised to enter pelvis. B, the specimen has been delivered. The levators are divided. C, separation of anterior attachments of the rectum. The prostate and membranous urethra are clearly seen.

39

Fig. 23.--Abdominoperineal resecClosure of the perineum completed, showing position of the soft rubber sump drain.

tlon.

is removed. Omentum is brought down and the perineum is closed tightly around a soft rubber sump drain approximating subcutaneous tissue and skin separately. The sump drain is inserted into the pelvis at the anterior end of the perineal incision upward and backward to place the tip in the presacral space (Fig. 23). Suction is applied immediately to remove blood before it can clot. Miles and his successors closed the pelvic peritoneum and placed large packs in the perineum to support the suture line, a practice that results in a large, rigid cavity that takes 3-6 months or longer to fill by granulation.* Others have reconstructed the pelvic floor and closed the perineum around drains. This step converts the pelvis into a dead space, the sutured peritoneum forming the roof and the perineum the floor. When the drain is removed, serum collect~ in the cavity and chronic infection is not uncommon. Closure of the pelvic peritoneum is invariably tenuous and often results in herniation of a loop of small intestine through the suture line and in intestinal obstruction. Since 1962 G7 our practice has been to suture only the skin and ischiorectal tissue of the perineum around a soft sump placed in the presacral space. The omentum and small bowel are allowed to fill the pelvic cavity, obliterating dead space. Infection is rare and the perineal wound heals per primam in 10-12 days. Regeneration of the parietal peritoneal lining of the pelvis is rapid, and intestinal obstruction is rare. Placement of the abdominal colostomy through the incision has the advantage of ease of performance. The colon traverses the abdominal cavity from back to front leaving no narrow lateral gutter to close. *No'rE: Since completion of this manuscript, a controlled clinical trial of three different methods of perineal wound management was published by T. T. Irvin and J. C. Goligher (Br. J. Surg. 62:287, 1975). The authors studied 106 patients undergoing excision of the rectum for inflammatory or malignant disease and demonstrated that 67.5% of perineal wounds treated by primary closure had healed completely within three months as opposed to only 30% of perineal wounds managed by the open method.

40

During closure of the laparotomy incision, the colon can be accurately sutured to" all layers under direct vision. This prevents herniation of small intestine or omentum alongside the stoma, as well as prolapse of the stoma itself. In our experience paracolostomy hernia rarely is seen following this method, in contrast to colostomy performed through a separate wound. Mucocutaneous suturing of the stoma minimizes scar formation and subsequent stenosis. Location of the stoma near the umbilicus is best for subsequent colostomy care, as a lower position may place it out of view on a pendulous abdomen.

REHABILITATION AFTER ABDOMINOPERINEALRESECTION It has been our experience that the rehabilitative process is more effective when under the direction of one individual--a nurse specialist, enterostomal therapist (E.T.), who has the expertise to serve as the resource person for the patient and his family during and after hospitalization as well as for the nursing, resident and surgical staffs. Placement of the stoma is crucial. The E.T. evaluates all parameters: accessibility and visibility, abdominal contours, especially deep fat rolls or depressions; location of the umbilicus; presence of old scars; belt level and bony prominences. With these factors in mind, the E.T. has the patient assume a variety of positions--lying, sitting, standing, bending--and thus determines the optimal location and marks the stoma site for the surgeon. Barring any severe or crippling infirmities, all patients are taught to irrigate approximately 7-10 days postoperatively. Management without irrigation has been neither successful nor satisfactory.54 We have found that 75% of patients who irrigate their colostomies experience control and wear nothing but a small gauze pad over the colostomy site. Two methods of irrigation are used: the enema technique and the bulb syringe method, s9 We advocate the bulb syringe method unless the patient is physically unable to master this technique. The first irrigation usually is given by the therapist on the fourth or fifth postoperative day with the patient seated on a commode or toilet. The practice of carrying out irrigation with the patient in bed is most undesirable and should be reserved only for those few patients who are bedridden. Following initial irrigation the patient is progressively taught the details of the method until he has learned to perform it himself. The patient should be self-sufficient in management of the irrigation before discharge from the hospital. Irrigation should be performed daily and at approximately the same time each day. Daily irrigations are more physiologically desirable: constipation, retention of gas, retention of irrigating fluid and other difficulties are found less frequently than when irrigations are performed at longer intervals of time. 41

Since ingestion of food is a stimulus to defecation, patients generally find irrigation more effective if performed after either breakfast or dinner. Ordinary tap water is used, usually at a tepid temperature and without additives. A few patients may have better results using cool water. Patients are on a regular diet upon discharge from the hospital and are encouraged to rely on past experience and trial and error with respect to foods that may cause distress; thus rigid, unnecessary and ritualistic restriction are avoided. There are no restrictions, other than those self-imposed by the patient, on sexual activity, travel, occupation or recreation. In patients who do not achieve control, the frequency and amount of spillage vary greatly. These patients continue to irrigate, by choice, but wear an odor-proof bag or pouch, changing it as necessary during the day. Generally those who fall into this category have had erratic bowel habits prior to operation.

SUMMARY The aim of surgery in the treatment of malignancies of the rectum is eradication of disease and restoration of the patient to a useful place in society. That this goal is not always achieved is demonstrated by the generally recognized over-all 5-year survival of no more than 4 5 % . Improvement in cure rate rests largely upon early detection of localized disease and new modalities for treatment of disseminated disease, but it is essential to emphasize that an appreciation of current knowledge of anatomy, pathology and physiology and the application of current methods of surgery and rehabilitation will achieve maximum benefit for each patient. Abdominoperineal resection and anterior resection are well-established methods of treatment of carcinoma o f the rectum. Study of the spread of cancer and the increasing experience with sphincter-saving TABLE I.--TREATMENT OF CARCINOMA OF THE RECTUM IN 229 PATIENTS---AGE, SEX AND OPERATIVE MORTALITY Aseo~mosAcrat Patients

XS~raXNOeEXXr~EX~. ANTERIOR ~ESEC'r~ON

50

40

139

Females Age

24 39-74 years mean: 60 years

12 49-86 years mean: 69 years

66 , 23-87 years mean: 62 years

Males Age

26 42-79 years mean: 60 years

28 40--75 years mean: 64 years

73 39-92 years mean: 64 years

Operative mortality

1 (2%)

1 (2.5%)

2 (1.4%)

42

operations indicate that sacrifice of the anus is not always essential for the cure o f rectal cancer. Our recent experience in treatment of carcinoma of the rectum by abdominosacral, abdominoperineal and anterior resection has been r e p o r t e d : s Data on sex, age and associated risk and operative mortality for this group of 229 patients, treated by one surgeon ( S A L ) , in a 6-year period are summarized in Table 1. The preponderance of males in the abdominoperineal group is not surprising since selection was based on level of lesion, and the wider female pelvis allows sphincter-saving procedures at a lower level. The three groups were otherwise comparable. The status of patients in this series in whom operation was considered curative is recorded in Table 2. A follow-up of up to 8 years shows that survival following abdominosacral resection compares favorably with abdominoperineal and anterior resection. Death from recurrence is correlated with stage of disease, not with level of lesion. Sphincter function following low anterior resection is essentially normal. Patients are continent of stool and flatus but, as alluded to in the section on continence, they initially experience a period of frequent small bowel movements due to loss of reservoir function. It is gratifying to note that functional results following abdominosacral resection differ only in degree from those observed following standard anterior resection. All patients were continent of stool and flatus from the outset and, after a somewhat longer period, some 12 weeks, of aberrant reservoir continence, all regained satisfactory function. This process of adaptation is facilitated by early institution of a diet high in fiber content and moisture, and encouragement of the patient to suppress the frequent urge to defecate. Patients in this series undergoing abdominosacral resection would under other circumsta0ces have TABLE 2.---SURVIVAL IN 195 PATIENTS HAVING CURATIVE OPERATIONS FOR 'RECTAL CARCINOMA

Patients Alive with recurrence Dead of recurrence Dead of other causes Alive and well

^Bt,OMINOSAC~XL 48*

*BDOMXNOPERtNr~L

37

ANrrRIORReSEC'nOS 110

1(2%) mean: 22 months

3(8.1%) mean: 35 months

5 (4.5%) mean: 28 months

6 (12.5%) mean: 22 months

12 (32%) mean: 27 months

26t (24%) mean: 15 months

4(8%)

1(2.5%)

36 (75%) mean: 32 months

21 (57%) mean: 27 months

*Includes operative death in one patient. 1"Includes 8 patients lost to follow-up.

43

5 (4.5%) 74 (64%) mean: 31 months

had abdominoperineal resection. A colostomy, even well constructed, properly managed and accepted by the patient, is at best a p o o r substitute for a functioning anal sphincter. Patients undergoing abdominoperineal resection with permanent sigmoid colostomy are visited by the enterostomal therapist preoperatively. Instruction in colostomy irrigation and management of colostomy appliances begins as soon as intestinal function returns. At discharge on the 14th to 21st postoperative day, patients are confident in colostomy care, and rehabilitation may be continued on an outpatient basis. In 2 - 3 months 7 5 % of patients achieve sufficient reservoir continence and regularity to spend the day without a colostomy appliance. Unless a patient is moribund from metastatic disease or has serious contraindications to operation due to severe intercurrent illness, an attempt at removal of the lesion should be made in every case. Little can be done to alleviate the suffering due to pain, tenesmus, infection and discharge of blood and mucus from an unrcsected rectal cancer. Indeed, the apparently inoperable growth may appear fixed because of its bulk rather than by invasion, and the surgical effort is then rewarded by effective palliation or cure.

ACKNOWLEDGMENTS We wish to thank Miss Ann Tobin for her inestimable editorial assistance and Miss Elizabeth Purcell, our enterostomal therapist, for her skill and useful information. REFERENCES 1. Ackerman, L. V., and Spratt, J. S.: Do adenomatous polyps become cancer?, Gastroenterology 44:905, 1963. 2. Allingham, W,, and Allingham, H.: The Diagnosis and Treatment o[ Diseases o] the Rectum (7th ed.; London: Baillere, 1901 ). 3. American Cancer Society: Cancer Facts and Figures, 1975 (New York: American Cancer Society, Inc., 1975). 4. Anderson, P. A., Dockerty, M. B., and Buie, L. A.: Myomatous tumors of the rectum (leiomyosarcomas and myosarcoma), Surgery 28:642, 1950. 5. Aries, V., Crowther, J. S., Drasar, B. S., Hill, M. J., and Williams, R. E. O.: Bacteria and aetiology of cancer of the large bowel, Gut 10:334, 1969. 6. Babcock, W. W.: The operative treatment of carcinoma of the rectosigmoid with methods for elimination of colostomy, Surg. Gynecol. Obstet. 55:627, 1932. 7. Bacon, H. E.: Abdominoperineal proctosigmoidectomy with sphincter preservation, J.A.M.A. 160:628, 1956. 8. Bacon, H. E.: Evolution of sphincter muscle preservation and re-establishment of continuity in the operative treatment of rectal and sigmoidal cancer, Surg. Gynecol. Obstet. 81 : 113, 1945. 9. Bacon, H. E.: Present status of the pull through sphincter saving procedure, Cancer 28:196, 1971. 10. Balfour, D. C.: A method of anastomosis between sigmoid and rectum, Am. Surg. 51:235, 1910. 11. Best, R. R., and Blair, J. B.: Sphincter preserving operations for rectal carcinoma as related to anatomy of lymphatics, Ann. Surg. 130:538, 1949. 44

12. Best, R. R., and Rasmussen, J. A.: Results of sphincter preserving operations for carcinoma of mid-rectum, J.A.M.A. 164:739, 1957. 13. Black, B. M.: Combined abdomino-endorectal resection. A surgical procedure preserxqng continuity of the bowel for the management of certain types of carcinoma of the mid rectum and upper part of the rectum, Mayo Clinic Proc. 23:545, 1948. 14. Black, B. M., and Botham, R. J.: Combined abdominoendorectal resection for lesions of the mid and upper parts of the rectum, Arch. Surg. 76:688, 1958. 15. Black, W. A., and Waugh, J. M.: The intramural extension of carcinoma of the descending colon, sigmoid and rectosigmoid, Surg. Gynec. Obstet. 87: 457, 1948. 16. Braastad, F. W., Dockerty, M. B., and Dixon, C. F.: Melanoepithelioma of the anus and rectum, Surgery 25:82, 1949. 17. Burkitt, D. P.: Epidemiology of cancer of the colon and rectum, Cancer 28:3, 1971. 18. Caldarola, V. T., Jackman, R. J., Moertel, C., and Dockerty, M. B.: Carcinoid tumors of the rectum, Am. J. Surg. 107:844, 1964. 19. Castleman, B., and Krichstein, H. I.: Do adenomatous polyps of the colon become malignant?, N. Engl. J. Med. 267:469, 1962. 20. Courtney, H.: Anatomy of the pelvic diaphragm and anorectal musculature as related to sphincter preservation in anorectal surgery, Am. J. Surg. 79:155, 1950. 21. Coller, F. A., Kay, E. B., and Maclntyre, R. S.: Regional lymphatic metastasis of carcinoma of the rectum, Surgery 8:294, 1940. 22. Crile, G., Jr., and Turnbull, R. B., Jr.: The role of electrocoagulation in the treatment of carcinoma of the rectum, Surg. Gynecol. Obstet. 135:391, 1972. 23. Cripps, H.: On Diseases of the R e c t u m and A n u s (4th cd.; London: Churchill, 1914). 24. Culp, C. E., and Hill, J. R.: Malignant lymphoma involving the rectum, Dis. Colon Rectum 5:426, 1962. 25. Cutait, D. E., and Figlione, F. J.: A new method of colorectal anastomosis, Dis. Colon Rectum 4:335, 1961. 26. Czerny, V., quoted in Rankin, F. W.: How surgery of the colon and rectum developed, Surg. Gynecol. Obstet. 64:705, 1937. 27. Dafoe, W. A.: The retrograde lymphatic spread of carcinoma of the rectum situated between the pectinate line and the peritoneal reflection, M. S. Thesis, University of Minnesota, 1944. 28. d'Allaines, F., and de Vernejour, R.: Conservation de la fonction sphincterienne dans la cure op6ratoire du cancer du rectum, J. Chir. (Paris) 62: 275, 1946. 29. Dawson, I. M. P., Comes, J. S., and Morson, B. C.: Primary malignant lymphoid tumours of the intestinal tract, Br. J. Surg. 49:80, 1961. 30. Dixon, C. F.: Anterior resection for carcinoma low in the sigmoid and the rectosigmoid, Surgery 15: 367, 1944. 31. Dixon, C. F.: Anterior resection for malignant lesions of the lower part of the rectum and lower part of the sigmoid, Ann. Surg. 128:425, 1948. 32. Dixon, C. F.: Surgical removal of lesions occurring in the sigmoid and rectosigmoid, Am. J. Surg. 46:12, 1939. 33. Donaldson, G. A., Rodkey, G. V., and Behringer, G. E.: Resection of the rectum with anal preservation, Surg. Gynecol. Obstet. 123:571, 1966. 34. Dukes, C.: An explanation of the difference between a papilloma and an adenoma of the rectum, Proc. R. Soe. l~led. 40:829, 1947. 35. Dukes, C. E.: Cancer of the rectum: An analysis of 1000 cases, J. Pathol. 50:527, 1932. 36. Dukes, C. E.: The classification of cancer of the rectum, J. Pathol. 35:323, 1932. 45

37. Duthie, H. L., and Gairns, F. W.: Sensory nerve endings and sensation in the anal region of man, Br. J. Surg. 47:585, 1960. 9 38. Duthie, H. L.: Progress report--Anal continence, Gut 12:844, 1971. 39. Ewing, M. R.: Villous tumors of the rectum, Ann. R. Coll. Surg.'Engl. 6:413, 1950. 40. Faget, J., quoted in Rankin, F. W.: How surgery of the colon and rectum developed, Surg. Gynecol. Obstet. 64:705, 1937. 41. Finsterer, H.: Zur chirurgischen Behandlung des Rectum-Carcinom, Arch. Klin. Chir. 202:15, 1941. 42. Gabriel, W. B.: The end result of perineal excision and of radium in treatment of cancer of the rectum, Br. J. Surg. 20:234, 1932. 43. Gabriel, W. B., Dukes, C., and Bussey, J. J. R.: The lymphatic spread in cancer of the rectum, Br. J. Surg. 23:395, 1935. 44. Gaston, E.: Physiological basis for preservation of fecal continence after resection of rectum, J.A.M.A. 146:1486, 1951. 45. Gilchrist, R. K., and David, V. C.: Lymphatic spread of carcinoma of the rectum, Ann. Surg. 108:621, 1938. 46. Glover, R. P., and Waugh, J. M.: The retrograde lymphatic spread of carcinoma of the rectosigmoid region, Surg. Gynecol. Obstet. 82:434, 1946. 47. Goetze, O.: Die abdominosacrale Resektion des Mastdarms mit wie der Hersttelung der naturlichen kon Kontinenz, Arch. Klin. Clair. 206:293, 1944. 48. Goligher, J. C.: The functional results after sphincter saving resection of the rectum, Ann. R. Coll. Surg. Engl. 8:421, 1951. 49. Goligher, J. C., Duthie, H. L., De Dombal, F. T., and Watts, J. McK.: Abdominoanal pull through excision for tumors of the mid third of the rectum, Br. J. Surg. 52:323, 1965. 50. Goligher, J. C., Graham, N. G., and De Dombal, F. T.: Anastomotic dehiscence after anterior resection of rectum and sigmoid, Br. J. Surg. 57:109, 1970. 51. Goligher, J. C., and Hughes, E. S. R.: Sensibility of the rectum and colon: Its role in the mechanism of anal continence, Lancet 270:543, 1951. 52. Goligher, J. C., Leacock, A. G., and Brossy, J. J.: The surgical anatomy of the anal canal, Br. J. Surg. 43:51, 1955. 53. Gosset, A., and Masson. P.: Tumeurs endocrines de l'appendice, Presse Med. 22:25, 1914. 54. Grief, W . R . N . , Postel, A. H., Syarse, A., and Localio, S. A.: An evaluation of colonic stoma management without irrigation, Surg. Gynecol. Obstet. 118:1234, 1964.. 55. Grinnell, R. S.: The lymphatic and venous spread of carcinoma of the rectum, Ann. Surg. 116:200, 1942. 56. Grinnell, R- S., and Lane, N.: Benign and malignant adenomatous polyps and papillary adenomas of colon and rectum: Analysis of 1856 tumors in 1335 patients, Int. Abst. Surg. 106:519, 1958. 57. Guernsey, D. F., Waugh, J. M., Dockerty, M. B.: Carcinoma of the rectum: Prognosis based on the distance of lesions from, or involvement of, the levator ani muscle, and involvement of the anal sphincters, Surg. Gynecol. Obstet. 92:529, 1951. 58. Helwig, E. B., and Hanson, J.: Lymphoid polyps (benign lymphoma) and malignant lymphoma of the rectum and anus, Surg. Gynecol. Obstet. 92: 233, 1951. 59. Hochenegg, J.: Meine Operation serfalge bei Rectum-Carcinom, Wien. Kiln. Wochenschr. 13:399, 1900. 60. Husa, A., and Hockerstedt, K.: Anorectal malignant melanoma, Acta Chir. Stand. 140:68, 1974. 61. Kligerman, M. M., and Urdanetta-Laffee, N.: Proceedings: Observations on 15 inoperable, non-resectable cases of rectal cancer given pre-op irradiation, Am. J. Roentgenol. Radium Ther. Nucl. Med. 120:624, 1974. 46

62. Kraske, P.: Zur Exstirpation hochsitzender Mastdarmkrebs, Verh. Dtsch. Ges. Chir. 14:464, 1885. 63. Lisfranc, J., quoted in Rankin, F. W.: How surgery of the colon and rectum developed, Surg. Gynecol. Obstet. 64:705, 1937. 64. Localio, S. A.: Abdominotransacral resection and anastomosis for mid rectal carcinoma, Surg. Gynecol. Obstet. 132:123, 1971. 65. Localio, S. A.: Curative surgery of mid-rectal cancer with preservation of the sphincters, in Nyhus, L. M. (ed.) Surgery ,4nnual (New York: Appleton-Century-Crofts, 1974), pp. 213-245. 66. Localio, S. A., and Baron, B.: Abdominotransacral resection and anastomosis for mid rectal cancer, Ann. Surg. 178:540, 1973. 67. Localio, S. A., Grier, W. R. N., and Postel, A. H.: Techniques of abdominoperineal resection of the rectum, Surg. Gynecol. Obstet. 129:598, 1967. 68. Lockhart-Mummery, J. P.: Diseases of the Rectum and Colon and Their Surg(eal Treatment (Baltimore: William Wood and Co., 1934). 69. Lockhart-Mummery, J. P.: Two hundred cases of cancer of the rectum treated by perineal excision, Br. J. Surg. 14:110, 1926. 70. Lubarsch, O.: Ueber den primaren Krebs des Ileum nebst Bemerkungen uber das gleichzeitige Vorkommen yon Krebs und Tuberculose, Virch. Arch. Pathol. Anat. I 11:280, 1888. 71. Madden, J. L., and Kandalaft, S.: Electrocoagulation: A primary and preferred method, Am. Surg. 166:413, 1967. 72. Madden, J. L., and Kandalaft, S.: Electrocoagulation in the treatment of cancer of the rectum. A continuing study, Am. Surg. 174:530, 1971. 73. Mandl, F.: Uber 1000 sakrale Mastdarmkreb Exstirpationem (darts der Hocheneggschen Material), Dtsch. Z. Chir. 219:3, 1929. 74. Maunsell, H. W.: A new method of excising the two upper portions of the rectum and the lower segment of the sigmoid flexure of the colon, Lancet 2:473, 1892. 75. McVay, J. R.: Involvement of lymph nodes in carcinoma of the rectum, Ann. Surg. 76:755, 1922. 76. Miles, W. E.: A method of performing abdomi~aoperineal excision for carcinoma of the rectum and of the terminal portion of the pelvic colon, Lancet 2:1812, 1908. 77. Miles, W. E.: Cancer of the Rectum (London: Harrison and Sons, Ltd., 1926). 78. Morson, B. C.: Polyps of the colon and rectum, in Ravitcb, M. (ed.) Current Problems hz Surgery (Chicago: Year Book Medical Publishers, Inc., 1970). 79. Morson, B. C.: Rare malignant tumors of the rectum and anal region, in Dukes, C. E. (ed.) Cancer of the Rectum (Edinburgh and London:E. & S. Livingstone, Ltd., 1960).. 80. blorson, B. C., and Dawson, I. M. P.: Gastrointestinal Pathology. (Oxford: Blackwell Scientific Publications, 1972). 81. Morson, B. C., and Volkstadt, H.: Malignant melanoma of the anal canal, J. Clin. Pathol. 16:126, 1963. 82. Oh, C., and Kark, A. E.: Anatomy of the external anal sphincter, Br. J. Surg. 59:717, 1972. 83. Pack, G. T., and Martins, F. G.: Treatment of anorectal malignant melanoma, Dis. Colon Rectum 3:15, 1960. 84. Papillon, J.: Resectable rectal cancers: Treatment by curative endocavitary irradiation, J.A.M.A. 231 : 1385, 1975. 85. Perry, P. M., Cross, R. M., and Morson, B. C.: Primary malignant lymphoma of the rectum, Proc. R. Soc. Med. 65:72, 1972. 86. Peskin, G. W., and Orloff, M. J.: A clinical study of 25 patients with carcinoid tumors of the rectum, Surg. Gynecol. Obstet. 109:673, 1959. 87. Postel, A. H., Grier, W. R. N., and Localio, S. A.: Simplified method of colostomy irrigation, Surg. Gynecol. Obstct. 121:595, 1965. 47

88. Quan, S. H. Q., Bader, G., and Berg, J. W.: Carcinoid tumors of the rectum, Dis. Colon Rectum 7:197, 1964. 89. Quan, S. H. Q., White, J. E., and Deddish, M. R.: Malignant melanoma of the anorectum, Dis. Colon Rectum 2:275, 1959. 90. Saegesser, F., and Gross, M.: Carcinoid syndrome and carcinoid tumors of the rectum, Am. J. Proctol. 20:27, 1969. 91. Saltykow, S. L.: Uber die Genese der kar-zlnoiden Tumoren sowie der Adenomyome des Darmes, Beitr. Pathol. Anat. 54:559, 1912. 92. SchSrli, A. F., and Kiesewetter, W. B.: Defecation and continence: Some new concepts, Dis. Colon Rectum 13:81, 1970. 93. Slanetz, C. A., Jr., and Herter, F. P.: The large intestine, in ttaagensen, C. D. (ed.) The Lymphatics in Cancer (Philadelphia: W. B. Saunders Co., 1972). 94. Spratt, J. S., Jr., Ackerman, L. V., and Moyer, C. A.: Relationship of polyps of colon to colonic cancer, Ann. Surg. 148:682, 1958. 95. Stearns, M. W., Jr., Deddish, M. R., and Quan, S. H. Q.: Roentgen therapy for cancer of the rectum, Surg. Gynecol. Obstet. 109:225, 1959. 96. Stearns, M. W., Jr., Deddish, M. R., Quan, S. H. Q., and Learning, R. H.: Pre-operative roentgen therapy for cancer of the rectum and rectosigmoid, Surg. Gynecol. Obstet. 138:584, 1974. 97. Stout, A. P.: Tumors of the colon and rectum, in Turell, I. R. (ed.) Diseases o] Colon and Rectum (2d ed.; Philadelphia: W. B. Saunders Co., 1969), Vol. 1. 98. Strauss, A. A., Strauss, S. F., Crawford, R. A., and Strauss, H. A.: Surgical diathermy of carcinoma of the rectum: Its clinical end results, J.A.M.A. 104:1480, 1935. 99. Strauss, A. A., Appel, M., Saphir, O., and Rabinovitz, A. J.: Immunological resistance to carcinoma produced by electrocoagulation, Surg. Gynecol. Obstet. 121:989, 1965. 100. Sunderland, D. A., and Binkley, G. E.: Papillary adenomas of the large intestine, Cancer 1: 184, 1948. 101. Turnbull, R. P., and Cuthbertson, A.: Abdominorectal pull-through for cancer and Hirschsprung's disease, delayed posterior colorectal anastomosis, Cleve. Clin. Q. 28:109, 1961. 102. Verneuil, A. A., quoted in Tuttle, J. P.: A Treatise on Diseases o[ the Anus, Rectum, and Pelvic Colon (2d ed.; New York: Appleton-Century-Crofts, 1905). 103. Villemin, F., Huard, P., and Montague, M.: Recherches anatomiques sur les lymphatiques du rectum et de l'anus, Rev. Chit. 63:39, 1925. 104. Waugh, J. M., Black, M. A., and Gage, R. P.: Three and five year survivals following combined abdominoperineal resection, abdominoperineal resection with sphincter preservation and anterior resection for carcinoma of the rectum and lower part of the sigmoid colon, Ann. Surg. 142:752, 1955. 105. Waugh, J. M., and Turner, J. C.: A study of 268 patients with carcinoma of the mid rectum treated by abdominoperineal resection with sphincter preservation, Surg. Gynecol. Obstet. 107:777, 1958. 106. Weir, R. F.: An improved method of treating high seated cancer of the rectum, J.A.M.A. 37:801, 1901. 107. Westhues, H.: Uber die Entstehung und Vermeidung des Iokalen RectumCarcinom-Rezidivs, Arch. Klin. Chit. 161:582, 1930. 108. Wheat, M. W., and Ackerman, L. V.: Villous adenomas of the large intestine, Ann. Surg. 147:476, 1958.

48