Journal of Pediatric Surgery 49 (2014) 534–538
Contents lists available at ScienceDirect
Journal of Pediatric Surgery journal homepage: www.elsevier.com/locate/jpedsurg
Pancreaticoduodenectomy for malignancies in children Giuseppe d'Ambrosio a, Laura del Prete a, Chiara Grimaldi a, Arianna Bertocchini a, Cristina Lo Zupone b, Lidia Monti b, Jean de Ville de Goyet a,⁎ a b
Department of Surgery and Transplantation Centre, Hepato-Biliary and Transplant Surgery Unit, Bambino Gesù Childrens Hospital, Rome, Italy Department of Imaging, Hepato-Biliary Radiology Unit, Bambino Gesù Childrens Hospital, Rome, Italy
a r t i c l e
i n f o
Article history: Received 9 June 2013 Received in revised form 28 July 2013 Accepted 9 September 2013 Key words: Pediatric malignancy Surgical indications Pancreaticoduodenectomy Portal vein reconstruction Outcome
a b s t r a c t Purpose: Malignant tumors of the common bile duct or of the pancreas head are uncommon in childhood [Perez EA, Gutierrez JC, Koniaris LG, Neville HL, Thompson WR, Sola JE. Malignant pancreatic tumors: incidence and outcome in 58 pediatric patients. J Pediatr Surg. 2009; Jan; 44 (1): 197–203]. With radical surgery being the standard cure for nonmetastatic diseases, pancreaticoduodenectomy (PD) is the best choice when the tumor is localized in the head of the pancreas, or in the lower portion of the common bile duct. The purpose of the present study is to describe five consecutive children managed by PD, and reviewing the particular aspects and results of this rare procedure in children. Methods: Between 2007 and 2010, five patients (median age: 7 years) underwent PD for nonmetastatic malignant tumors. In two cases, PD was performed en bloc with a right hepatectomy in order to achieve the radical resection of a recurrent biliary sarcoma. Four patients benefited from a “pylorus-preserving” PD procedure. In two patients, resection of the portal vein and vascular reconstruction was performed, and in one case, an extended resection of the biliary ductal system was necessary. Results: All resection margins were clear. The postoperative course was uneventful, with no pancreatic or biliary leakage in all of the patients. Oral refeeding was achieved by the eighth postoperative day. In two cases, a late revision of pancreatic–jejunal anastomosis was performed because of mild steatorrea and a suspected anastomotic stricture. Two of the patients, who were subsequently operated on second hand, for biliary sarcoma, died from the recurrence; while three of the others, with pancreatic malignancies, are alive and well, with a good functional outcome. Conclusions: Surgical resection is the treatment of choice for tumors of the pancreatic head area. In the absence of regional or metastatic extension, the radicality of primary intervention is associated with favorable outcomes. Good functionality results were observed after the PD was limited to the head of the pancreas and subject to pyloruspreserving techniques. © 2014 Elsevier Inc. All rights reserved.
In contrast to adult patients in whom the disease is relatively frequent, pancreatic and common bile duct neoplasms are utterly uncommon in the pediatric population. The pancreatic cancer incidence is 0.2% of all pediatric malignancies [1], with among these, pancreatoblastoma being typical of early childhood, while solid or cystic pseudopapillary tumors and neuroendocrine neoplasms can be observed in older groups [2]. In the literature, pancreatic carcinomas and sarcomas are linked to the worst prognosis [2,3]. PD was first performed successfully by Kausch in 1909, and subsequently diffused by Whipple in 1935 [4]. Owing to the rarity of pancreatic cancer in children, it is a procedure that is very rarely performed in this age group. The aim of this retrospective study is to report our experience with PD, its indication, and its results thereof.
⁎ Corresponding author. Bambino Gesù Childrens Hospital Piazza Sant'Onofrio 4, 00165 Rome, Italy. Tel.: +39 0668592851; fax: +39 0668593841. E-mail address: [email protected] (J. de Ville de Goyet). 0022-3468/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpedsurg.2013.09.010
1. Patients and methods 1.1. Patients A retrospective review of the hospital database, for a 4-year study period (2007 to 2010), identified five patients who underwent a PD procedure. Indications were a pancreatic neoplasm in three cases (two pancreatic acinar cell carcinomas, one solid pseudopapillary tumor—ultrasound-guided needle biopsy diagnostic) and rhabdomyosarcoma of the bile ducts (RMSB) in two patients. The latter two patients (Table 1: cases D and E) had been diagnosed, biopsied and initially managed in another center, and they were referred after adjuvant chemotherapy; their locoregional anatomy was complex, and an en bloc DP + partial hepatectomy was considered to be necessary in order to achieve a radical resection (none had metastases, either at diagnosis or preoperatively). In one of these two cases (case D), an explorative laparotomy had been performed at presentation (for biliary obstruction with pseudocystic transformation of the extrahepatic bile ducts), and where a surgical exploration of the “cyst,” a biopsy and drainage had been performed: the biopsy
G. d'Ambrosio et al. / Journal of Pediatric Surgery 49 (2014) 534–538
535
Table 1 Demographics and characteristics of five children submitted to pancreaticoduodenectomy. Gender
Age at surgery (years)
Case A
M
7
Case B
F
6.5
Case C
F
12
Case D
F
3
Case E
M
6.5
Types of tumor
Surgery
Complications after surgery
Follow-up (months)
Pancreatic acinar cell carcinoma (Fig. 3) Pancreatic acinar cell carcinoma Solid pseudopapillary tumor
Pylorus-sparing PD + portal vein resection and reconstruction Pylorus-sparing PD
–
26 41
Rhabdomyosarcoma of the extra hepatic biliary tract Rhabdomyosarcoma of the extrahepatic biliary tract (Fig. 2)
Pylorus-sparing PD + right hepatectomy PD + conversion to right trisegmentectomy + portal vein reconstruction
Stenosis of pancreatic–jejunal anastomosis/resolved Stenosis of pancreatic–jejunal anastomosis/resolved – –
12 (death)
Pylorus-sparing PD
confirmed the diagnosis of RMSB, but the procedure was further complicated by a chronic biliary fistula to the wound. Although chemotherapy reduced much of the diameter of the tumor mass, the anatomy of the tumor, the previous surgery, and the (likely) neoplastic fistula, would have made it difficult to achieve a radical resection by conventional (limited) resection. The other patient (case E) had previously undergone a right hepatectomy for RMSB in another center, which was followed by a recurrence of the tumor in and around the liver hilum and along the extrahepatic bile duct, which was closely attached to the portal vein trunk. 1.2. Technique From a technical point of view, the PD was performed according to the classical procedure, as is well described in the literature, but was
24 24 (death)
modified with the preservation of the pylorus when appropriate [4,5] (Fig. 1). The pylorus was preserved when, in the absence of abnormal lymph nodes (at preoperative imaging or at surgery), along the common hepatic artery, a free margin of a minimum of 3 cm was found (macroscopically and during the intervention) between the pylorus and the tumor mass. In two cases, the PD was performed en bloc with the hepatectomy, after the identification and the preservation of the left hepatic artery, and of the portal vein, and without dividing the liver hilum; in one of the two patients (case E), a portion of the portal vein trunk was in continuity with the specimen. When the resection of the tumor mass had been completed, reconstruction of the digestive system was performed as follows: A Roux jejunal loop was fashioned and positioned through the Treitz hiatus, to reach the supramesocolic area, where it was anastomosed in an end-to-end manner to the pancreatic stump, by separate stitches of
Fig. 1. Technical steps for digestive system reconstruction after pancreaticoduodenectomy (A to D). (A) Pancreaticojejunal end-to-end anastomosis, in double layers (interrupted prolene) with intussusception of the pancreatic stump (2) to cover the first layer (1) (with or without internal stent) (R: Roux-en-Y loop). (B) The hepaticojejunal anastomosis is performed in second, end-to-side, with continuous absorbable monofilament suture and without stent. (C) The 50-cm Jejunal Roux loop has been routed in the supramesocolic area through the Treitz hiatus (T). It is positioned in order to be anastomosed with the gastric outlet (or pylorus) in an isoperistaltic and antireflux (to the Roux) manner. (MC: transverse mesocolon). (D) The gastric outlet—or pylorus (P)—is brought through the mesocolon in order to bring the gastrojejunal anastomosis (continuous absorbable monofilament) within the inframesocolic abdominal area, so that the latter anastomosis is anatomically distant from the previous ones (S: stomach).
536
G. d'Ambrosio et al. / Journal of Pediatric Surgery 49 (2014) 534–538
prolene; this anastomotic line was covered (with intussusception of the pancreas stump) by a similar layer. The bile duct was anastomosed end-to-side to the Roux loop, 10 cm downstream, by running a 6/0 monofilament absorbable suture. The reconstruction of the gastric outlet consisted of an end-to-side anastomosis between the pylorus, or the gastric antrum, to the loop: in order to separate this anastomoses from the previous ones, the distal portion of the stomach was brought through the mesocolon, within the inframesocolic area, where it was anastomosed end-to-side to the loop, 50 cm down to the biliopancreatic. 1.3. Methods Outcomes were evaluated as an occurrence, or not, of early perioperative complications (first hospital stay), late complications, and the need of reoperations, and by assessing the functional outcomes (glucose tolerance measured by fasted glucose, 24-hour blood glucose, and malabsorption, by measuring the level of fat and elastase in the stools), growth, recurrence rate and survival (Table 1). 2. Results In all cases, the whole of the pancreatic head was resected, together with part of the body in two cases that presented very large pancreatic masses (cases A and C). A “pylorus-preserving” PD was performed in four patients; while in one case, the pylorus and part of the stomach were resected in continuity with PD for achieving radicality (case E—a large RMSB that also necessitated a partial portal vein resection). In two RMSB patients (cases D and E), the PD was performed with en-bloc resection of the porta hepatis, and either the right liver (N = 1, case D), or the remnant of the right liver with the left anterior sector of the liver in the other case (N = 1, case E). In these two cases, the extrahepatic biliary structures were entirely removed, including the biliary confluence and the hilar plate: the proximal division line was at the level of the left bile duct in one case, and at the level of the segmental ducts (segments II and III) in the second case (Fig. 2). In two cases (cases A and E), the portal vein trunk had to be partially resected for technical and oncological reasons (in one case, with the resection of the splenomesenteric confluence). Vascular reconstruction consisted of direct end-to-end anastomosis (portal
vein to the splenomesenteric confluence in one patient, and the left portal vein to the portal trunk in another case) (Figs. 2 and 3). Postoperative recovery was uneventful in all of the patients, with none of the complications commonly associated with PD, such as pancreatic, biliary or digestive fistulas, vascular thrombosis, or postoperative bleeding, being observed this series. All patients were initiated postoperatively with parenteral nutrition, but oral feeding was restarted by the first week, and was completed by the end of the second week [four patients were out of parenteral support at postoperative day 11, with only one necessitating support for 3 weeks (case A)]. Interestingly, preserving the pylorus was not associated with gastric-emptying problems or “dumping syndrome.” At mid- and long-term points, no postoperative ulcer, abdominal discomfort, chronic vomiting, or diarrhea was observed. In three children, supplementation by pancreatic enzymes was necessary during the first year, including the two children who underwent the larger pancreatic resection for oncological reasons. The first of the latter two children (case A—the one who had the largest resection— Fig. 3) also needed temporary insulin supportive therapy (the first 2 weeks after surgery): glucose regulation was checked later (for fasted glycemia and glycemia Holter) and was turned back to normal, but he developed an abnormal ultrasonographic imaging of the liver, typical of hepatic steatosis, during the first 4 postoperative months. Currently, 4 years after surgery, he is under a normal diet, with minimal pancreatic enzymes support, and growing nicely with normal fasted glucose, no clinical signs of malabsorption, and a normal ultrasonographic appearance of the liver. The two other cases were suspected of having a stricture of the pancreaticojejunal anastomosis (ultrasound and cholangio RM), and were reoperated (at 8 and 12 months after the first surgery, respectively): the reintervention simply consisted of opening the anterior aspect of the pancreaticojejunal anastomosis, with resection of the pancreas margin (sent for histological review—both negative), identification of the dilated Wirsung duct, and pancreaticojejunal reanastomosis on an internal stent. This was followed by a significant improvement of the intestinal absorption (as assessed clinically, and by the normalization of the elastase test of stools). However, one of these two cases (case C), who had normal glucose metabolism for the first year after surgery, at the 3-year follow-up, displayed slightly higher than normal fast glycemia, was overweight, and had a typical
Fig. 2. A patient treated for biliary sarcoma: extrahepatic biliary structures were entirely removed (including the biliary confluence and the left bile duct) in continuity with all the surrounding hilar tissue and head of the pancreas: portal vein was partially resected. The two tubes are in bile ducts of hepatic segment II and segment III, respectively (case E—Table 1).
G. d'Ambrosio et al. / Journal of Pediatric Surgery 49 (2014) 534–538
537
Fig. 3. A patient treated for huge pancreatic acinar cell carcinoma of the head of the pancreas. The lower portion of the portal vein trunk had to be resected with the splenomesenteric confluence, for technical and oncological reasons. Vascular reconstruction consisted of direct end-to-end anastomosis (case A—Table 1).
profile of a bad eating lifestyle, with poor physical activity. This patient is currently under continued observation. The median follow-up was 24 months (range: 12 to 41 months) (Table 1). The two patients operated on second hand for biliary sarcoma died from a recurrence of the disease, at 12 and 24 months, postoperatively, respectively. The other three are alive, and have been free of disease for 24 to 41 months after their operations, and are growing nicely on a normal diet and within the 50–90 percentile range. 3. Discussion As an intervention, PD is rarely performed in children, since the indication is mostly limited to complex trauma cases or malignancies involving the head of the pancreas, both of which are extremely rare. This “rarity” explains that published series, and the number of cases in each report, are a limited resource in the literature [2]. This short series is reported to contribute to the knowledge about managing these patients and for analyzing the feasibility and the outcome of PD in children. Pancreatic tumors in childhood are most commonly presented with a palpable mass, or with pain, and in the absence of metastases, as commonly, these tumors have an expansive, rather than, an infiltrative growth pattern [6]. Radical surgical excision is the gold standard of management, even at the price of aggressive resections, as this is associated with a good prognosis and of survival [2,3,7]. However, it is not rare that some teams choose a relatively conservative approach when it comes to operating on children (local and partial resections with dubious margins). Although the reluctance to perform aggressive surgery in this age group is understandable, this approach is associated with a higher incidence of recurrence, local or metastatic. Our experience, with a more
aggressive radical approach, is limited, but suggests that this type of surgery is providing a radicality of the resection, and is associated with low rates of morbidity and mortality, when appropriately indicated and performed by experienced teams. Although RMSB is a neoplasm of the bile ducts, and possibly can extend all along the common bile duct downstream, there has been no previous report of a combined en bloc hepatectomy–PD for RMSB in children: we report two cases for whom this strategy was proposed, eventually as a second choice instead of primary surgery, because of a complex tumoral anatomy, or because of recurrence. Both children were carefully assessed, and an aggressive resection of the malignant area seemed to be appropriate management, although it necessitated an en bloc resection of the liver, together with all extrahepatic biliary structures and surrounding tissue. Considering that radical surgery is key in the management of a good outcome, it seems logical to propose aggressive surgery, when no other option is technically possible [8]. Because both cases were “second hand,” with a complex local anatomical condition, it is difficult to comment on what (of the former or of the latter) was the worst factor leading to the recurrence and the bad outcome (cases D and E). Although RMSB will remain an unusual indication for PD, and independently of the outcome of the two patients in this series, we consider that such an approach may be justified in the future, as primary surgery, in selected patients, in order to offer a complete resection. As with any complex intervention, a PD is associated with a significant risk of technical complications, late morbidity, and even mortality, and these have been well described in the adult literature. The rate of mortality and morbidity, associated with the implementation of a PD in adult patients, is approximately 1%–3% and 30%–40%, respectively [9]: biliopancreatic fistula, gastric or jejunal fistulas, sepsis, and various types of digestive dysfunctions. Fistula of the
538
G. d'Ambrosio et al. / Journal of Pediatric Surgery 49 (2014) 534–538
pancreatic–jejunal anastomosis is the most common, with a 2% to 20% occurrence rate [10]. Technical refinements, as proposed, for example, by Yang et al. [9] (terminoterminal pancreaticojejunal anastomosis with a continuous double layer suture), have helped to prevent this complication in adults; it is difficult to implement that technique in children, whose pancreas is smaller in thickness/diameter and having a soft parenchyma. In this series, interrupted suturing was used, and the absence of complications in the five cases suggests that careful and usual suturing techniques be used by pediatric surgeons and is associated with good healing and results. Another important element when performing a PD in children is to consider the possibility of preserving the pylorus in order to preserve gastric functioning. “Pylorus-sparing PD” was reported by Traverso and Longemire [11] in 1978. Although it is still matter of debate in the literature, a “pylorus-sparing” procedure is associated with an improved postoperative gastrointestinal function, and a reduction of the incidence of both a dumping syndrome, and ulcerations of the anastomosis, and with an improved postoperative nutritional status [5]. In four of our five patients, it was possible to preserve the pyloric valve. In one case only, it was impossible to spare the pylorus, because the extension of the neoplasm required a wider excision to achieve radicality. In all five children, it was possible to restart oral nutrition, early after the intervention (6th to 8th postoperative day). At midand long-term, none of our patients presented gastric-emptying disorders, or other digestive dysmotility. In our series of patients, there were no statistically significant differences when compared to what is reported in the literature with regard to operative time and hospital days [4]. In the absence of a large pediatric series in the literature, it is difficult to make comparisons and draw rules. This series, however, confirms previous observations, such as, for example, a female/male ratio of 3:2 [1,12], and that longterm survival depends mostly on the histological type and the possibility of obtaining a radical resection [7]. Importantly, this short series has focused the attention on the technical aspects of the operation, and the adequate preservation of pancreatic functions in the long term, showing that a PD can be proposed safely for children, if appropriate attention is given to the operation, and to the preservation of the healthy adjacent structures, maximizing the preservation
of a normal pancreas and the stomach. Vascular reconstruction can be safely performed when necessary. 4. Conclusions Overall, this series shows that a PD is associated with good results in children, both in terms of the radical excision of pancreatic (nonmetastatic) neoplasms and in terms of (the low rate of) morbidity and mortality. Importantly, long-term results in the three surviving children suggest that, although a large part of the pancreatic mass has been resected, endocrine and exocrine functions can be relatively well preserved, and that normal growth can be maintained. References [1] Perez EA, Gutierrez JC, Koniaris LG, et al. Malignant pancreatic tumors: incidence and outcome in 58 pediatric patients. J Pediatr Surg 2009;44(1):197–203. [2] Jaksic T, Yaman M, Thorner P, et al. A 20 year review of pediatric pancreatic tumors. J Pediatr Surg 1992;27:1315–7. [3] Grosfeld JL, Vane DW, Rescorla FJ, et al. Pancreatic tumors in childhood: analysis of 13 cases. J Pediatr Surg 1990;25:1057–62. [4] Cameron JL, Riall TS, Coleman J, et al. One thousand consecutive pancreaticoduodenectomies. Ann Surg 2006;244:10–5. [5] Kennedy EP, Brumbaugh J, Yeo CJ. Reconstruction following the pylorus preserving Whipple resection: PJ, HJ, and DJ. J Gastrointest Surg 2010;14(2):408–15. [6] Shorter NA, Glick RD, Klimstra DS, et al. Malignant pancreatic tumors in childhood and adolescence: the Memorial Sloan–Kettering experience, 1967 to present. J Pediatr Surg 2002;37:887–92. [7] Rojas Y, Warneke CL, Dhamne CA, et al. Primary malignant pancreatic neoplasms in children and adolescents: a 20 year experience. J Pediatr Surg 2012 Dec;47(12):2199–204. [8] Wang KS, Albanese C, Dada F, et al. Papillary cystic neoplasm of the pancreas: a report of three pediatric cases and literature review. J Pediatr Surg 1998 Jun;33(6):842–5. [9] Yang YL, Xu XP, Wu GQ, et al. Prevention of pancreatic leakage after pancreaticoduodenectomy by modified child pancreaticojejunostomy. Hepatobiliary Pancreat Dis Int 2008 Aug;7(4):426–9. [10] Abete M, Ronchetti V, Casano A, et al. Pancreatic fistula after pancreaticoduodenectomy: risk factors and treatment. Minerva Chir 2005;60:99–110. [11] Traverso LW, Longmire Jr WP. Preservation of the pylorus in pancreaticoduodenectomy a follow-up evaluation. Ann Surg 1980 Sep;192(3):306–10. [12] Park J, Dunn JC, Atkinson JB. Management of children with a pancreatic head mass. J Pediatr Surg 2006;41:1–4.
Contents lists available at ScienceDirect
Journal of Pediatric Surgery journal homepage: www.elsevier.com/locate/jpedsurg
Pancreaticoduodenectomy for malignancies in children Giuseppe d'Ambrosio a, Laura del Prete a, Chiara Grimaldi a, Arianna Bertocchini a, Cristina Lo Zupone b, Lidia Monti b, Jean de Ville de Goyet a,⁎ a b
Department of Surgery and Transplantation Centre, Hepato-Biliary and Transplant Surgery Unit, Bambino Gesù Childrens Hospital, Rome, Italy Department of Imaging, Hepato-Biliary Radiology Unit, Bambino Gesù Childrens Hospital, Rome, Italy
a r t i c l e
i n f o
Article history: Received 9 June 2013 Received in revised form 28 July 2013 Accepted 9 September 2013 Key words: Pediatric malignancy Surgical indications Pancreaticoduodenectomy Portal vein reconstruction Outcome
a b s t r a c t Purpose: Malignant tumors of the common bile duct or of the pancreas head are uncommon in childhood [Perez EA, Gutierrez JC, Koniaris LG, Neville HL, Thompson WR, Sola JE. Malignant pancreatic tumors: incidence and outcome in 58 pediatric patients. J Pediatr Surg. 2009; Jan; 44 (1): 197–203]. With radical surgery being the standard cure for nonmetastatic diseases, pancreaticoduodenectomy (PD) is the best choice when the tumor is localized in the head of the pancreas, or in the lower portion of the common bile duct. The purpose of the present study is to describe five consecutive children managed by PD, and reviewing the particular aspects and results of this rare procedure in children. Methods: Between 2007 and 2010, five patients (median age: 7 years) underwent PD for nonmetastatic malignant tumors. In two cases, PD was performed en bloc with a right hepatectomy in order to achieve the radical resection of a recurrent biliary sarcoma. Four patients benefited from a “pylorus-preserving” PD procedure. In two patients, resection of the portal vein and vascular reconstruction was performed, and in one case, an extended resection of the biliary ductal system was necessary. Results: All resection margins were clear. The postoperative course was uneventful, with no pancreatic or biliary leakage in all of the patients. Oral refeeding was achieved by the eighth postoperative day. In two cases, a late revision of pancreatic–jejunal anastomosis was performed because of mild steatorrea and a suspected anastomotic stricture. Two of the patients, who were subsequently operated on second hand, for biliary sarcoma, died from the recurrence; while three of the others, with pancreatic malignancies, are alive and well, with a good functional outcome. Conclusions: Surgical resection is the treatment of choice for tumors of the pancreatic head area. In the absence of regional or metastatic extension, the radicality of primary intervention is associated with favorable outcomes. Good functionality results were observed after the PD was limited to the head of the pancreas and subject to pyloruspreserving techniques. © 2014 Elsevier Inc. All rights reserved.
In contrast to adult patients in whom the disease is relatively frequent, pancreatic and common bile duct neoplasms are utterly uncommon in the pediatric population. The pancreatic cancer incidence is 0.2% of all pediatric malignancies [1], with among these, pancreatoblastoma being typical of early childhood, while solid or cystic pseudopapillary tumors and neuroendocrine neoplasms can be observed in older groups [2]. In the literature, pancreatic carcinomas and sarcomas are linked to the worst prognosis [2,3]. PD was first performed successfully by Kausch in 1909, and subsequently diffused by Whipple in 1935 [4]. Owing to the rarity of pancreatic cancer in children, it is a procedure that is very rarely performed in this age group. The aim of this retrospective study is to report our experience with PD, its indication, and its results thereof.
⁎ Corresponding author. Bambino Gesù Childrens Hospital Piazza Sant'Onofrio 4, 00165 Rome, Italy. Tel.: +39 0668592851; fax: +39 0668593841. E-mail address: [email protected] (J. de Ville de Goyet). 0022-3468/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpedsurg.2013.09.010
1. Patients and methods 1.1. Patients A retrospective review of the hospital database, for a 4-year study period (2007 to 2010), identified five patients who underwent a PD procedure. Indications were a pancreatic neoplasm in three cases (two pancreatic acinar cell carcinomas, one solid pseudopapillary tumor—ultrasound-guided needle biopsy diagnostic) and rhabdomyosarcoma of the bile ducts (RMSB) in two patients. The latter two patients (Table 1: cases D and E) had been diagnosed, biopsied and initially managed in another center, and they were referred after adjuvant chemotherapy; their locoregional anatomy was complex, and an en bloc DP + partial hepatectomy was considered to be necessary in order to achieve a radical resection (none had metastases, either at diagnosis or preoperatively). In one of these two cases (case D), an explorative laparotomy had been performed at presentation (for biliary obstruction with pseudocystic transformation of the extrahepatic bile ducts), and where a surgical exploration of the “cyst,” a biopsy and drainage had been performed: the biopsy
G. d'Ambrosio et al. / Journal of Pediatric Surgery 49 (2014) 534–538
535
Table 1 Demographics and characteristics of five children submitted to pancreaticoduodenectomy. Gender
Age at surgery (years)
Case A
M
7
Case B
F
6.5
Case C
F
12
Case D
F
3
Case E
M
6.5
Types of tumor
Surgery
Complications after surgery
Follow-up (months)
Pancreatic acinar cell carcinoma (Fig. 3) Pancreatic acinar cell carcinoma Solid pseudopapillary tumor
Pylorus-sparing PD + portal vein resection and reconstruction Pylorus-sparing PD
–
26 41
Rhabdomyosarcoma of the extra hepatic biliary tract Rhabdomyosarcoma of the extrahepatic biliary tract (Fig. 2)
Pylorus-sparing PD + right hepatectomy PD + conversion to right trisegmentectomy + portal vein reconstruction
Stenosis of pancreatic–jejunal anastomosis/resolved Stenosis of pancreatic–jejunal anastomosis/resolved – –
12 (death)
Pylorus-sparing PD
confirmed the diagnosis of RMSB, but the procedure was further complicated by a chronic biliary fistula to the wound. Although chemotherapy reduced much of the diameter of the tumor mass, the anatomy of the tumor, the previous surgery, and the (likely) neoplastic fistula, would have made it difficult to achieve a radical resection by conventional (limited) resection. The other patient (case E) had previously undergone a right hepatectomy for RMSB in another center, which was followed by a recurrence of the tumor in and around the liver hilum and along the extrahepatic bile duct, which was closely attached to the portal vein trunk. 1.2. Technique From a technical point of view, the PD was performed according to the classical procedure, as is well described in the literature, but was
24 24 (death)
modified with the preservation of the pylorus when appropriate [4,5] (Fig. 1). The pylorus was preserved when, in the absence of abnormal lymph nodes (at preoperative imaging or at surgery), along the common hepatic artery, a free margin of a minimum of 3 cm was found (macroscopically and during the intervention) between the pylorus and the tumor mass. In two cases, the PD was performed en bloc with the hepatectomy, after the identification and the preservation of the left hepatic artery, and of the portal vein, and without dividing the liver hilum; in one of the two patients (case E), a portion of the portal vein trunk was in continuity with the specimen. When the resection of the tumor mass had been completed, reconstruction of the digestive system was performed as follows: A Roux jejunal loop was fashioned and positioned through the Treitz hiatus, to reach the supramesocolic area, where it was anastomosed in an end-to-end manner to the pancreatic stump, by separate stitches of
Fig. 1. Technical steps for digestive system reconstruction after pancreaticoduodenectomy (A to D). (A) Pancreaticojejunal end-to-end anastomosis, in double layers (interrupted prolene) with intussusception of the pancreatic stump (2) to cover the first layer (1) (with or without internal stent) (R: Roux-en-Y loop). (B) The hepaticojejunal anastomosis is performed in second, end-to-side, with continuous absorbable monofilament suture and without stent. (C) The 50-cm Jejunal Roux loop has been routed in the supramesocolic area through the Treitz hiatus (T). It is positioned in order to be anastomosed with the gastric outlet (or pylorus) in an isoperistaltic and antireflux (to the Roux) manner. (MC: transverse mesocolon). (D) The gastric outlet—or pylorus (P)—is brought through the mesocolon in order to bring the gastrojejunal anastomosis (continuous absorbable monofilament) within the inframesocolic abdominal area, so that the latter anastomosis is anatomically distant from the previous ones (S: stomach).
536
G. d'Ambrosio et al. / Journal of Pediatric Surgery 49 (2014) 534–538
prolene; this anastomotic line was covered (with intussusception of the pancreas stump) by a similar layer. The bile duct was anastomosed end-to-side to the Roux loop, 10 cm downstream, by running a 6/0 monofilament absorbable suture. The reconstruction of the gastric outlet consisted of an end-to-side anastomosis between the pylorus, or the gastric antrum, to the loop: in order to separate this anastomoses from the previous ones, the distal portion of the stomach was brought through the mesocolon, within the inframesocolic area, where it was anastomosed end-to-side to the loop, 50 cm down to the biliopancreatic. 1.3. Methods Outcomes were evaluated as an occurrence, or not, of early perioperative complications (first hospital stay), late complications, and the need of reoperations, and by assessing the functional outcomes (glucose tolerance measured by fasted glucose, 24-hour blood glucose, and malabsorption, by measuring the level of fat and elastase in the stools), growth, recurrence rate and survival (Table 1). 2. Results In all cases, the whole of the pancreatic head was resected, together with part of the body in two cases that presented very large pancreatic masses (cases A and C). A “pylorus-preserving” PD was performed in four patients; while in one case, the pylorus and part of the stomach were resected in continuity with PD for achieving radicality (case E—a large RMSB that also necessitated a partial portal vein resection). In two RMSB patients (cases D and E), the PD was performed with en-bloc resection of the porta hepatis, and either the right liver (N = 1, case D), or the remnant of the right liver with the left anterior sector of the liver in the other case (N = 1, case E). In these two cases, the extrahepatic biliary structures were entirely removed, including the biliary confluence and the hilar plate: the proximal division line was at the level of the left bile duct in one case, and at the level of the segmental ducts (segments II and III) in the second case (Fig. 2). In two cases (cases A and E), the portal vein trunk had to be partially resected for technical and oncological reasons (in one case, with the resection of the splenomesenteric confluence). Vascular reconstruction consisted of direct end-to-end anastomosis (portal
vein to the splenomesenteric confluence in one patient, and the left portal vein to the portal trunk in another case) (Figs. 2 and 3). Postoperative recovery was uneventful in all of the patients, with none of the complications commonly associated with PD, such as pancreatic, biliary or digestive fistulas, vascular thrombosis, or postoperative bleeding, being observed this series. All patients were initiated postoperatively with parenteral nutrition, but oral feeding was restarted by the first week, and was completed by the end of the second week [four patients were out of parenteral support at postoperative day 11, with only one necessitating support for 3 weeks (case A)]. Interestingly, preserving the pylorus was not associated with gastric-emptying problems or “dumping syndrome.” At mid- and long-term points, no postoperative ulcer, abdominal discomfort, chronic vomiting, or diarrhea was observed. In three children, supplementation by pancreatic enzymes was necessary during the first year, including the two children who underwent the larger pancreatic resection for oncological reasons. The first of the latter two children (case A—the one who had the largest resection— Fig. 3) also needed temporary insulin supportive therapy (the first 2 weeks after surgery): glucose regulation was checked later (for fasted glycemia and glycemia Holter) and was turned back to normal, but he developed an abnormal ultrasonographic imaging of the liver, typical of hepatic steatosis, during the first 4 postoperative months. Currently, 4 years after surgery, he is under a normal diet, with minimal pancreatic enzymes support, and growing nicely with normal fasted glucose, no clinical signs of malabsorption, and a normal ultrasonographic appearance of the liver. The two other cases were suspected of having a stricture of the pancreaticojejunal anastomosis (ultrasound and cholangio RM), and were reoperated (at 8 and 12 months after the first surgery, respectively): the reintervention simply consisted of opening the anterior aspect of the pancreaticojejunal anastomosis, with resection of the pancreas margin (sent for histological review—both negative), identification of the dilated Wirsung duct, and pancreaticojejunal reanastomosis on an internal stent. This was followed by a significant improvement of the intestinal absorption (as assessed clinically, and by the normalization of the elastase test of stools). However, one of these two cases (case C), who had normal glucose metabolism for the first year after surgery, at the 3-year follow-up, displayed slightly higher than normal fast glycemia, was overweight, and had a typical
Fig. 2. A patient treated for biliary sarcoma: extrahepatic biliary structures were entirely removed (including the biliary confluence and the left bile duct) in continuity with all the surrounding hilar tissue and head of the pancreas: portal vein was partially resected. The two tubes are in bile ducts of hepatic segment II and segment III, respectively (case E—Table 1).
G. d'Ambrosio et al. / Journal of Pediatric Surgery 49 (2014) 534–538
537
Fig. 3. A patient treated for huge pancreatic acinar cell carcinoma of the head of the pancreas. The lower portion of the portal vein trunk had to be resected with the splenomesenteric confluence, for technical and oncological reasons. Vascular reconstruction consisted of direct end-to-end anastomosis (case A—Table 1).
profile of a bad eating lifestyle, with poor physical activity. This patient is currently under continued observation. The median follow-up was 24 months (range: 12 to 41 months) (Table 1). The two patients operated on second hand for biliary sarcoma died from a recurrence of the disease, at 12 and 24 months, postoperatively, respectively. The other three are alive, and have been free of disease for 24 to 41 months after their operations, and are growing nicely on a normal diet and within the 50–90 percentile range. 3. Discussion As an intervention, PD is rarely performed in children, since the indication is mostly limited to complex trauma cases or malignancies involving the head of the pancreas, both of which are extremely rare. This “rarity” explains that published series, and the number of cases in each report, are a limited resource in the literature [2]. This short series is reported to contribute to the knowledge about managing these patients and for analyzing the feasibility and the outcome of PD in children. Pancreatic tumors in childhood are most commonly presented with a palpable mass, or with pain, and in the absence of metastases, as commonly, these tumors have an expansive, rather than, an infiltrative growth pattern [6]. Radical surgical excision is the gold standard of management, even at the price of aggressive resections, as this is associated with a good prognosis and of survival [2,3,7]. However, it is not rare that some teams choose a relatively conservative approach when it comes to operating on children (local and partial resections with dubious margins). Although the reluctance to perform aggressive surgery in this age group is understandable, this approach is associated with a higher incidence of recurrence, local or metastatic. Our experience, with a more
aggressive radical approach, is limited, but suggests that this type of surgery is providing a radicality of the resection, and is associated with low rates of morbidity and mortality, when appropriately indicated and performed by experienced teams. Although RMSB is a neoplasm of the bile ducts, and possibly can extend all along the common bile duct downstream, there has been no previous report of a combined en bloc hepatectomy–PD for RMSB in children: we report two cases for whom this strategy was proposed, eventually as a second choice instead of primary surgery, because of a complex tumoral anatomy, or because of recurrence. Both children were carefully assessed, and an aggressive resection of the malignant area seemed to be appropriate management, although it necessitated an en bloc resection of the liver, together with all extrahepatic biliary structures and surrounding tissue. Considering that radical surgery is key in the management of a good outcome, it seems logical to propose aggressive surgery, when no other option is technically possible [8]. Because both cases were “second hand,” with a complex local anatomical condition, it is difficult to comment on what (of the former or of the latter) was the worst factor leading to the recurrence and the bad outcome (cases D and E). Although RMSB will remain an unusual indication for PD, and independently of the outcome of the two patients in this series, we consider that such an approach may be justified in the future, as primary surgery, in selected patients, in order to offer a complete resection. As with any complex intervention, a PD is associated with a significant risk of technical complications, late morbidity, and even mortality, and these have been well described in the adult literature. The rate of mortality and morbidity, associated with the implementation of a PD in adult patients, is approximately 1%–3% and 30%–40%, respectively [9]: biliopancreatic fistula, gastric or jejunal fistulas, sepsis, and various types of digestive dysfunctions. Fistula of the
538
G. d'Ambrosio et al. / Journal of Pediatric Surgery 49 (2014) 534–538
pancreatic–jejunal anastomosis is the most common, with a 2% to 20% occurrence rate [10]. Technical refinements, as proposed, for example, by Yang et al. [9] (terminoterminal pancreaticojejunal anastomosis with a continuous double layer suture), have helped to prevent this complication in adults; it is difficult to implement that technique in children, whose pancreas is smaller in thickness/diameter and having a soft parenchyma. In this series, interrupted suturing was used, and the absence of complications in the five cases suggests that careful and usual suturing techniques be used by pediatric surgeons and is associated with good healing and results. Another important element when performing a PD in children is to consider the possibility of preserving the pylorus in order to preserve gastric functioning. “Pylorus-sparing PD” was reported by Traverso and Longemire [11] in 1978. Although it is still matter of debate in the literature, a “pylorus-sparing” procedure is associated with an improved postoperative gastrointestinal function, and a reduction of the incidence of both a dumping syndrome, and ulcerations of the anastomosis, and with an improved postoperative nutritional status [5]. In four of our five patients, it was possible to preserve the pyloric valve. In one case only, it was impossible to spare the pylorus, because the extension of the neoplasm required a wider excision to achieve radicality. In all five children, it was possible to restart oral nutrition, early after the intervention (6th to 8th postoperative day). At midand long-term, none of our patients presented gastric-emptying disorders, or other digestive dysmotility. In our series of patients, there were no statistically significant differences when compared to what is reported in the literature with regard to operative time and hospital days [4]. In the absence of a large pediatric series in the literature, it is difficult to make comparisons and draw rules. This series, however, confirms previous observations, such as, for example, a female/male ratio of 3:2 [1,12], and that longterm survival depends mostly on the histological type and the possibility of obtaining a radical resection [7]. Importantly, this short series has focused the attention on the technical aspects of the operation, and the adequate preservation of pancreatic functions in the long term, showing that a PD can be proposed safely for children, if appropriate attention is given to the operation, and to the preservation of the healthy adjacent structures, maximizing the preservation
of a normal pancreas and the stomach. Vascular reconstruction can be safely performed when necessary. 4. Conclusions Overall, this series shows that a PD is associated with good results in children, both in terms of the radical excision of pancreatic (nonmetastatic) neoplasms and in terms of (the low rate of) morbidity and mortality. Importantly, long-term results in the three surviving children suggest that, although a large part of the pancreatic mass has been resected, endocrine and exocrine functions can be relatively well preserved, and that normal growth can be maintained. References [1] Perez EA, Gutierrez JC, Koniaris LG, et al. Malignant pancreatic tumors: incidence and outcome in 58 pediatric patients. J Pediatr Surg 2009;44(1):197–203. [2] Jaksic T, Yaman M, Thorner P, et al. A 20 year review of pediatric pancreatic tumors. J Pediatr Surg 1992;27:1315–7. [3] Grosfeld JL, Vane DW, Rescorla FJ, et al. Pancreatic tumors in childhood: analysis of 13 cases. J Pediatr Surg 1990;25:1057–62. [4] Cameron JL, Riall TS, Coleman J, et al. One thousand consecutive pancreaticoduodenectomies. Ann Surg 2006;244:10–5. [5] Kennedy EP, Brumbaugh J, Yeo CJ. Reconstruction following the pylorus preserving Whipple resection: PJ, HJ, and DJ. J Gastrointest Surg 2010;14(2):408–15. [6] Shorter NA, Glick RD, Klimstra DS, et al. Malignant pancreatic tumors in childhood and adolescence: the Memorial Sloan–Kettering experience, 1967 to present. J Pediatr Surg 2002;37:887–92. [7] Rojas Y, Warneke CL, Dhamne CA, et al. Primary malignant pancreatic neoplasms in children and adolescents: a 20 year experience. J Pediatr Surg 2012 Dec;47(12):2199–204. [8] Wang KS, Albanese C, Dada F, et al. Papillary cystic neoplasm of the pancreas: a report of three pediatric cases and literature review. J Pediatr Surg 1998 Jun;33(6):842–5. [9] Yang YL, Xu XP, Wu GQ, et al. Prevention of pancreatic leakage after pancreaticoduodenectomy by modified child pancreaticojejunostomy. Hepatobiliary Pancreat Dis Int 2008 Aug;7(4):426–9. [10] Abete M, Ronchetti V, Casano A, et al. Pancreatic fistula after pancreaticoduodenectomy: risk factors and treatment. Minerva Chir 2005;60:99–110. [11] Traverso LW, Longmire Jr WP. Preservation of the pylorus in pancreaticoduodenectomy a follow-up evaluation. Ann Surg 1980 Sep;192(3):306–10. [12] Park J, Dunn JC, Atkinson JB. Management of children with a pancreatic head mass. J Pediatr Surg 2006;41:1–4.
