Outcome of salvage hepatic resection for recurrent hepatocellular carcinoma after radiofrequency ablation therapy Suguru Yamashita, MD, PhD, Taku Aoki, MD, PhD, Yosuke Inoue, MD, PhD, Junichi Kaneko, MD, PhD, Yoshihiro Sakamoto, MD, PhD, Yasuhiko Sugawara, MD, PhD, Kiyoshi Hasegawa, MD, PhD, and Norihiro Kokudo, MD, PhD, Bunkyo-ku, Tokyo, Japan
Background. Although radiofrequency ablation (RFA) is an effective local ablative technique for the treatment of hepatocellular carcinoma (HCC), the optimal treatment for recurrence after RFA has not been established. Methods. Between September 2002 and December 2011, 46 hepatectomies (salvage group) were performed for intrahepatic (local or multifocal) recurrent HCC after RFA. The difference between the imaging findings before RFA and at the time of salvage resection, especially in the Local recurrent group, and the short-term and long-term outcomes after salvage surgery were analyzed retrospectively by comparing them with those for a matched control group (n = 46) and with those of patients who underwent a second hepatic resection for HCC recurrence after an initial hepatic resection during the same period (n = 155). Results. The tumor-occupying region was more distributed widely before the salvage resection compared with that before RFA, and a more extensive operation would have been required (rather than the RFA) in the local group. An evaluation of the short-term outcomes revealed that salvage resection required a longer operative time and was associated with a greater frequency of morbidity. The long-term outcomes of the salvage group were poorer than those of patients who underwent repeat hepatic resection for HCC recurrence after an initial hepatic resection. Conclusion. The indications for RFA should be determined carefully, because recurrence after RFA may be associated with a more aggressive pattern of recurrence, and the long-term results after salvage resection are unsatisfactory. (Surgery 2015;157:463-72.) From the Hepato-Biliary-Pancreatic Surgery Division, Department of Surgery, Graduate School of Medicine, University of Tokyo, Bunkyo-ku, Tokyo, Japan
HEPATOCELLULAR CARCINOMA (HCC) is the fifth most common malignancy worldwide and causes nearly 1 million deaths each year. Hepatic resection, liver transplantation, and local forms of ablative therapy have been recognized as curative treatments for HCC. Among them, ultrasound-guided radiofrequency ablation (RFA) has proven to be a safe and effective therapy for small-sized HCC based on several randomized controlled trials and Accepted for publication October 31, 2014. Reprint requests: Norihiro Kokudo, MD, PhD, Hepato-BiliaryPancreatic Surgery Division, Department of Surgery, Graduate School of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan. E-mail: KOKUDO-2SU@h. u-tokyo.ac.jp. 0039-6060/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2014.10.019
systematic reviews.1-6 Recurrence after RFA, however, reportedly is common,7 and the optimal treatment after recurrence has not yet been established. Although repeated RFA has been a common treatment of choice after recurrence,8 few reports have documented the outcome of repeated RFA for recurrent HCC after RFA.9,10 Some recurrent lesions exhibit aggressive recurrence patterns,11-14 and salvage hepatic resection can be considered in such cases. To our knowledge, reports on the short-term and long-term outcomes of patients undergoing salvage resection for recurrent HCC after RFA are limited.15-20 Thus, the aim of the present study was to evaluate the technical aspects and the outcomes of salvage resection for recurrent HCC after RFA at our institution. The technical aspects were elucidated by focusing on 2 groups: a group with local recurrences after RFA, and a group with multifocal SURGERY 463
464 Yamashita et al
Surgery March 2015
Fig 1. Representative case showing an aggressive local recurrence after RFA and changes in the distribution of resection types in the local group (n = 30) before RFA and at the time of salvage resection. (A) A 67-year-old man who was infected with hepatitis B virus. An HCC nodule, 2 cm in diameter, was found in segment 8 near the anterior superior branch of the portal vein (white arrowhead). No findings of vascular invasion were observed. (B) The patient received RFA as an initial treatment. A CT scan performed 4 days after RFA showed an RFA scar in segment 8 (white arrow). (C) Two months after RFA, a local HCC recurrence 5 cm in diameter (white asterisk) with portal invasion in the anterior segment (white arrowhead) was observed. An anterior segmentectomy was performed for the recurrent HCC as a salvage resection. (D) The tumor-occupying region before, at time of salvage operation, and actual resection performed.
recurrences at sites other than the formerly treated sites. The short-term outcomes of the patients were then compared with those of a historically matched control group. As for the long-term outcomes, the survival benefit of salvage resection was analyzed by comparing survival after salvage resection to patients who underwent a second hepatic resection for HCC recurrence after an initial hepatic resection during the same period at the same institution. PATIENTS AND METHODS Patients. Between September 2002 and December 2011, 974 hepatic resections were performed for HCC at our institution. Among them, salvage resection was performed for intrahepatic, recurrent HCC after RFA in 46 consecutive patients. The RFA was not performed by a surgeon but by a gastroenterological physician at our institution or at another local institution using a percutaneous, transhepatic approach guided by ultrasonography. These patients comprised the study population (Salvage group) and were evaluated.
The number, size, and location of the tumors both before RFA and before salvage resection were evaluated via abdominal ultrasonography, computed tomography (CT), and ethoxy benzyl diethylenetriamine pentaacetic acid-enhanced magnetic resonance imaging. Salvage resection was performed for local recurrence after RFA in 30 patients (local group; Fig 1, A–C) and for multifocal recurrence other than at the formerly treated sites in the remaining 16 patients (35%; multifocal group). The indications for salvage resection were classified into the following categories: technical difficulty of repeated RFA, tumor thrombus, patient preference, and coexistence of localized peritoneal dissemination near the liver. Hepatic resection. Basically, our selection criteria for salvage resection were the same as those used for primary resection.21-23 In summary, the indications for hepatic resection and the types of operative procedures were judged after an evaluation of CT volumetric images and the liver functional reserve. Anatomic resection has been regarded as the preferred resection method for
Surgery Volume 157, Number 3
HCC, and the hepatectomies were performed as described previously.21,24 Localized peritoneal dissemination near the liver was radically resected. In all the patients in the present study, we attempted to perform a macroscopically curative salvage resection based on preoperative imaging results. The exceptions were cases in which a portal vein tumor thrombus (PVTT) extended beyond the bifurcation or into other sectors. In such cases, the PVTT was dissected intraluminally from the portal venous wall and removed through the venotomy, and macroscopically residual PVTTs protruding into tiny branches were meticulously extracted (peeling-off technique).25 Patients were excluded from salvage resection if they had extrahepatic metastasis, extended peritoneal dissemination, advanced cirrhosis, or severe comorbidity. When dissecting adhesions, a policy of preserving the adjacent organs was adopted, rather than performing an extended resection combined with adjacent organs. Especially in the local group, the type of operative procedure deemed to have been necessary if RFA had not been chosen was compared with the actual procedure used for salvage resection. After operation, the resected specimens were examined pathologically, and the tumor staging was determined based on the Japanese Tumor-Node-Metastasis staging system.26 All postoperative complications occurring within 2 months after salvage or repeat hepatectomy were graded according to the Clavien-Dindo classification.27-29 To assess the short-term influence of the RFA procedure on the intraoperative and postoperative findings, the operation time, blood loss, rate of R0 resection rate, and rate of postoperative complications among the patients who received salvage resections were obtained and compared with those in 46 patients who underwent hepatectomy as the initial treatment for HCC (matched for type of resection, sex, age, hepatitis etiology, indocyanine green retention rate at 15 minutes, Child Pugh Score, tumor status, and follow-up period: matched control group) during the same period. To assess the long-term influence of the RFA procedure, the long-term outcome of the salvage resection group was compared with those of patients who underwent a second hepatic resection for HCC recurrence after the initial hepatic resection during the same period at the same institution (n = 155). Follow-up after hepatectomy. All the patients were followed up with a clinical examination, measurement of the alpha-fetoprotein level, and
Yamashita et al 465
abdominal ultrasonography performed at 1 month after the resection and at 3-month intervals thereafter. Enhanced CT imaging was performed every 6 months.24,30 Recurrence was defined as the appearance of a new lesion with radiologic features typical of HCC. Recurrence-free survival was defined as the interval between the operation and the date of the diagnosis of the first recurrence or the last follow-up. Statistical analysis. Continuous data are reported as the median (range). Quantitative and categorized variables were compared using the Wilcoxon rank-sum test and the v2 test, respectively. The Kaplan-Meier method was used to calculate the survival rates, which were compared using a log rank test. A multivariable analysis of the prognostic survival factors was performed using a Cox regression model. The statistical analysis was performed using JMP software (version 9.0.0; SAS Institute Inc, Cary, NC). RESULTS Patient characteristics (Table I). As for the treatments performed before salvage resection, 32 of the 46 patients (70%) had undergone transcatheter arterial chemoembolization treatment at least once. Between the recurrence after RFA and the salvage resection, selective transcatheter arterial chemoembolization was performed in 11 patients as a preoperative treatment in an attempt to decrease tumor thrombi. As for the tumor status before RFA, the median tumor number was 1 (range, 1–3), and the median tumor size was 26 mm (range, 10–35). All the cases met the Milan selection criteria,31 and the use of RFA seemed to be appropriate. Characteristics of the salvage resections (Table I). Among the 46 resections, 30 hepatic resections (65%) were performed for local progression of the RFA-treated lesion (local group). The extent of the hepatic resection was less than 1 Couinaud segment in 16 patients (35%), 1 Couinaud segment in 10 patients (22%), 2 Couinaud segments in 13 patients (28%), and 3 or more Couinaud segments in 7 patients (15%). In 18 patients (39%), an anatomic resection could not be performed because of extensive adhesions in the post-RFA region. In 8 patients (17%), the operations resulted in microscopically noncurative resections (R0/R1/R2, 38/8/0). In each of the 6 R1 cases, the operative margin was pathologically positive at the plane of transaction at the post-RFA scar area. In the remaining 2 cases, a peeling-off technique was adopted to remove macroscopic PVTT; thus, the tumor was exposed
466 Yamashita et al
Surgery March 2015
Table I. Clinical characteristics of patients in Salvage group n = 46 Age, y* Sex (M:F) Hepatitis B/C/non-B, non-C Number of treatments before salvage resection TACE, n = 32 TAI, n = 1 PEIT, n = 9 ICGR15 at hepatectomy, %* Albumin, g/dL* Total bilirubin, mg/dL* Prothrombin, %* Child Pugh Score, 5/6/7 AFP, ng/mL* Local progression of the RFA-treated lesion, n Reasons for salvage hepatic resection, n Technical difficulty of repeat RFA Tumor thrombus Patient’s preference Coexistence of peritoneal dissemination Difficulty of operative procedure, n Extensive adhesion to diaphragm Extensive adhesion to adjacent organ other than diaphragm Impairment of adjacent major hepatic vein Extensive adhesion to intrathoracic space Operation time, min* Blood loss, mL* $1,000 Usage of RBC transfusion, n Usage of fresh frozen plasma transfusion, n No residual cancer, n R0/R1/R2 Anatomic resection, n Deviation from preoperative planning, n Extended procedure Less-invasive procedure No. of tumors* Maximal size of tumors, mm* Tumor status, n Portal invasion Venous invasion Bile duct invasion Intrahepatic metastasis Grade: well/moderate/poor/ unknown
67 (42–81) 42:4 9/34/3
1 0 0 15 3.9 0.7 88.1
(0–15) (0–1) (0–5) (4.8–45) (2.7–4.7) (0.2–1.8) (63–100) 30/15/1 9.95 (2–18,425) 30 (65%)
27 13 4 2
(59%) (28%) (9%) (4%)
28 (61%) 20 (44%) 5 (11%) 2 (4%) 1 (2%) 424 860 19 3 28
(98–810) (100–3,545) (41%) (7%) (61%)
38 (83%) 38/8/0 28 (61%) 38 (83%) 22 16 1 39
(48%) (35%) (1–11) (10–110)
23 (50%) 7 (15%) 2 (4%) 3 (7%) 5/29/8/4 (continued)
Table I. (continued) n = 46 Noncancerous tissue CH/LC, n Pathological stage I/II/III/IVA, n Morbidity, ny Hospital stay, d*
22/24 4/12/20/10 12 (26%) 16 (7–49)
*Median (range). yComplications of Clavien-Dindo grade $3a. Values in parentheses are percentages, unless indicated otherwise. AFP, Alpha-fetoprotein; CH, chronic hepatitis; ICGR15, indocyanine green retention rate at 15 min; LC, liver cirrhosis; PEIT, percutaneous ethanol injection therapy; RBC, red blood cells; RFA, radiofrequency ablation; TACE, transcatheter arterial chemoembolization; TAI, transcatheter arterial infusion.
grossly and was completely extracted. As for the pathologic findings, microscopic portal venous and hepatic venous tumor invasion were observed in 23 patients (50%) and 7 patients (15%), respectively. Moderately differentiated HCC was most common (n = 29, 63%), whereas well-differentiated HCC rarely was observed (n = 5, 11%). Characteristics of the local and multifocal groups. A comparison of the local and multifocal groups (Table II) revealed that an extended operative procedure was required more frequently in the local group (20/30, 67%), compared with the multifocal group (2/16, 13%) (P < .01). The intraoperative blood loss was greater in the local group than in the multifocal group (1,140 [180–3,110] mL and 435 [100–3,545] mL, respectively; P < .01). The local group required a longer duration of operation (443 [235–783] minutes), compared with the multifocal group (311 [98–810] minutes) (P = .057). In the local group, the median maximum size of the tumors was 27 mm (range, 15–35) before RFA and was 43 mm (range, 10–110) before salvage resection (Table II). Twenty-four patients (80%) exhibited an increase in the size of the local recurrence after RFA. Occult tumor thrombi were noted in 9 cases (30%) before the salvage resection, whereas no tumor thrombi were observed prior to RFA. The tumor-occupying region was distributed more widely before salvage resection than before RFA. The number of cases requiring a resection involving a greater number of Couinaud segments was greater before salvage resection than before RFA (Fig 1, D). Twenty patients (67%) actually required operations that were more extensive than the operations that would have been required before RFA. Although most of the patients (28/30, 93%) fulfilled the commonly adopted criteria for
Yamashita et al 467
Surgery Volume 157, Number 3
Table II. Comparison of operative characteristics between local and multifocal groups Local group (n = 30) Operation time, miny Blood loss, mLy Red blood cell transfusion, mLy Fresh frozen plasma transfusion, mLy Anatomic resection Deviation from the presumed type of resection Extended procedure Less-invasive procedure No residual cancer (R0) No. of tumorsy Maximum size of tumors, mmy Tumor status Portal invasion Venous invasion Bile duct invasion Intrahepatic metastasis Grade: well/moderate/poor/unknown Noncancerous tissue CH/LC Morbidityx Biliary leakage Accumulation of pleural or ascitic fluid Mortality Hospital stay, dy
443 1,140 0 800 24 24 20 4 24 1 43
(235–783) (180–3,110) (0–520) (0–5,040) (80) (80) (67) (13) (80) (1–9) (10–110)
17 (57) 6 (20) 2 (7) 2 (7) 4/16/7/3 15/15 8 (27) 3 (10) 5 (17) 0 (0) 16 (11–44)
Multifocal group (n = 16) 311 435 0 0 4 14 2 12 14 1 26
P value*
(98–810) (100–3,545) (0–0) (0–3,200) (25) (88)
.057z
Background. Although radiofrequency ablation (RFA) is an effective local ablative technique for the treatment of hepatocellular carcinoma (HCC), the optimal treatment for recurrence after RFA has not been established. Methods. Between September 2002 and December 2011, 46 hepatectomies (salvage group) were performed for intrahepatic (local or multifocal) recurrent HCC after RFA. The difference between the imaging findings before RFA and at the time of salvage resection, especially in the Local recurrent group, and the short-term and long-term outcomes after salvage surgery were analyzed retrospectively by comparing them with those for a matched control group (n = 46) and with those of patients who underwent a second hepatic resection for HCC recurrence after an initial hepatic resection during the same period (n = 155). Results. The tumor-occupying region was more distributed widely before the salvage resection compared with that before RFA, and a more extensive operation would have been required (rather than the RFA) in the local group. An evaluation of the short-term outcomes revealed that salvage resection required a longer operative time and was associated with a greater frequency of morbidity. The long-term outcomes of the salvage group were poorer than those of patients who underwent repeat hepatic resection for HCC recurrence after an initial hepatic resection. Conclusion. The indications for RFA should be determined carefully, because recurrence after RFA may be associated with a more aggressive pattern of recurrence, and the long-term results after salvage resection are unsatisfactory. (Surgery 2015;157:463-72.) From the Hepato-Biliary-Pancreatic Surgery Division, Department of Surgery, Graduate School of Medicine, University of Tokyo, Bunkyo-ku, Tokyo, Japan
HEPATOCELLULAR CARCINOMA (HCC) is the fifth most common malignancy worldwide and causes nearly 1 million deaths each year. Hepatic resection, liver transplantation, and local forms of ablative therapy have been recognized as curative treatments for HCC. Among them, ultrasound-guided radiofrequency ablation (RFA) has proven to be a safe and effective therapy for small-sized HCC based on several randomized controlled trials and Accepted for publication October 31, 2014. Reprint requests: Norihiro Kokudo, MD, PhD, Hepato-BiliaryPancreatic Surgery Division, Department of Surgery, Graduate School of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan. E-mail: KOKUDO-2SU@h. u-tokyo.ac.jp. 0039-6060/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2014.10.019
systematic reviews.1-6 Recurrence after RFA, however, reportedly is common,7 and the optimal treatment after recurrence has not yet been established. Although repeated RFA has been a common treatment of choice after recurrence,8 few reports have documented the outcome of repeated RFA for recurrent HCC after RFA.9,10 Some recurrent lesions exhibit aggressive recurrence patterns,11-14 and salvage hepatic resection can be considered in such cases. To our knowledge, reports on the short-term and long-term outcomes of patients undergoing salvage resection for recurrent HCC after RFA are limited.15-20 Thus, the aim of the present study was to evaluate the technical aspects and the outcomes of salvage resection for recurrent HCC after RFA at our institution. The technical aspects were elucidated by focusing on 2 groups: a group with local recurrences after RFA, and a group with multifocal SURGERY 463
464 Yamashita et al
Surgery March 2015
Fig 1. Representative case showing an aggressive local recurrence after RFA and changes in the distribution of resection types in the local group (n = 30) before RFA and at the time of salvage resection. (A) A 67-year-old man who was infected with hepatitis B virus. An HCC nodule, 2 cm in diameter, was found in segment 8 near the anterior superior branch of the portal vein (white arrowhead). No findings of vascular invasion were observed. (B) The patient received RFA as an initial treatment. A CT scan performed 4 days after RFA showed an RFA scar in segment 8 (white arrow). (C) Two months after RFA, a local HCC recurrence 5 cm in diameter (white asterisk) with portal invasion in the anterior segment (white arrowhead) was observed. An anterior segmentectomy was performed for the recurrent HCC as a salvage resection. (D) The tumor-occupying region before, at time of salvage operation, and actual resection performed.
recurrences at sites other than the formerly treated sites. The short-term outcomes of the patients were then compared with those of a historically matched control group. As for the long-term outcomes, the survival benefit of salvage resection was analyzed by comparing survival after salvage resection to patients who underwent a second hepatic resection for HCC recurrence after an initial hepatic resection during the same period at the same institution. PATIENTS AND METHODS Patients. Between September 2002 and December 2011, 974 hepatic resections were performed for HCC at our institution. Among them, salvage resection was performed for intrahepatic, recurrent HCC after RFA in 46 consecutive patients. The RFA was not performed by a surgeon but by a gastroenterological physician at our institution or at another local institution using a percutaneous, transhepatic approach guided by ultrasonography. These patients comprised the study population (Salvage group) and were evaluated.
The number, size, and location of the tumors both before RFA and before salvage resection were evaluated via abdominal ultrasonography, computed tomography (CT), and ethoxy benzyl diethylenetriamine pentaacetic acid-enhanced magnetic resonance imaging. Salvage resection was performed for local recurrence after RFA in 30 patients (local group; Fig 1, A–C) and for multifocal recurrence other than at the formerly treated sites in the remaining 16 patients (35%; multifocal group). The indications for salvage resection were classified into the following categories: technical difficulty of repeated RFA, tumor thrombus, patient preference, and coexistence of localized peritoneal dissemination near the liver. Hepatic resection. Basically, our selection criteria for salvage resection were the same as those used for primary resection.21-23 In summary, the indications for hepatic resection and the types of operative procedures were judged after an evaluation of CT volumetric images and the liver functional reserve. Anatomic resection has been regarded as the preferred resection method for
Surgery Volume 157, Number 3
HCC, and the hepatectomies were performed as described previously.21,24 Localized peritoneal dissemination near the liver was radically resected. In all the patients in the present study, we attempted to perform a macroscopically curative salvage resection based on preoperative imaging results. The exceptions were cases in which a portal vein tumor thrombus (PVTT) extended beyond the bifurcation or into other sectors. In such cases, the PVTT was dissected intraluminally from the portal venous wall and removed through the venotomy, and macroscopically residual PVTTs protruding into tiny branches were meticulously extracted (peeling-off technique).25 Patients were excluded from salvage resection if they had extrahepatic metastasis, extended peritoneal dissemination, advanced cirrhosis, or severe comorbidity. When dissecting adhesions, a policy of preserving the adjacent organs was adopted, rather than performing an extended resection combined with adjacent organs. Especially in the local group, the type of operative procedure deemed to have been necessary if RFA had not been chosen was compared with the actual procedure used for salvage resection. After operation, the resected specimens were examined pathologically, and the tumor staging was determined based on the Japanese Tumor-Node-Metastasis staging system.26 All postoperative complications occurring within 2 months after salvage or repeat hepatectomy were graded according to the Clavien-Dindo classification.27-29 To assess the short-term influence of the RFA procedure on the intraoperative and postoperative findings, the operation time, blood loss, rate of R0 resection rate, and rate of postoperative complications among the patients who received salvage resections were obtained and compared with those in 46 patients who underwent hepatectomy as the initial treatment for HCC (matched for type of resection, sex, age, hepatitis etiology, indocyanine green retention rate at 15 minutes, Child Pugh Score, tumor status, and follow-up period: matched control group) during the same period. To assess the long-term influence of the RFA procedure, the long-term outcome of the salvage resection group was compared with those of patients who underwent a second hepatic resection for HCC recurrence after the initial hepatic resection during the same period at the same institution (n = 155). Follow-up after hepatectomy. All the patients were followed up with a clinical examination, measurement of the alpha-fetoprotein level, and
Yamashita et al 465
abdominal ultrasonography performed at 1 month after the resection and at 3-month intervals thereafter. Enhanced CT imaging was performed every 6 months.24,30 Recurrence was defined as the appearance of a new lesion with radiologic features typical of HCC. Recurrence-free survival was defined as the interval between the operation and the date of the diagnosis of the first recurrence or the last follow-up. Statistical analysis. Continuous data are reported as the median (range). Quantitative and categorized variables were compared using the Wilcoxon rank-sum test and the v2 test, respectively. The Kaplan-Meier method was used to calculate the survival rates, which were compared using a log rank test. A multivariable analysis of the prognostic survival factors was performed using a Cox regression model. The statistical analysis was performed using JMP software (version 9.0.0; SAS Institute Inc, Cary, NC). RESULTS Patient characteristics (Table I). As for the treatments performed before salvage resection, 32 of the 46 patients (70%) had undergone transcatheter arterial chemoembolization treatment at least once. Between the recurrence after RFA and the salvage resection, selective transcatheter arterial chemoembolization was performed in 11 patients as a preoperative treatment in an attempt to decrease tumor thrombi. As for the tumor status before RFA, the median tumor number was 1 (range, 1–3), and the median tumor size was 26 mm (range, 10–35). All the cases met the Milan selection criteria,31 and the use of RFA seemed to be appropriate. Characteristics of the salvage resections (Table I). Among the 46 resections, 30 hepatic resections (65%) were performed for local progression of the RFA-treated lesion (local group). The extent of the hepatic resection was less than 1 Couinaud segment in 16 patients (35%), 1 Couinaud segment in 10 patients (22%), 2 Couinaud segments in 13 patients (28%), and 3 or more Couinaud segments in 7 patients (15%). In 18 patients (39%), an anatomic resection could not be performed because of extensive adhesions in the post-RFA region. In 8 patients (17%), the operations resulted in microscopically noncurative resections (R0/R1/R2, 38/8/0). In each of the 6 R1 cases, the operative margin was pathologically positive at the plane of transaction at the post-RFA scar area. In the remaining 2 cases, a peeling-off technique was adopted to remove macroscopic PVTT; thus, the tumor was exposed
466 Yamashita et al
Surgery March 2015
Table I. Clinical characteristics of patients in Salvage group n = 46 Age, y* Sex (M:F) Hepatitis B/C/non-B, non-C Number of treatments before salvage resection TACE, n = 32 TAI, n = 1 PEIT, n = 9 ICGR15 at hepatectomy, %* Albumin, g/dL* Total bilirubin, mg/dL* Prothrombin, %* Child Pugh Score, 5/6/7 AFP, ng/mL* Local progression of the RFA-treated lesion, n Reasons for salvage hepatic resection, n Technical difficulty of repeat RFA Tumor thrombus Patient’s preference Coexistence of peritoneal dissemination Difficulty of operative procedure, n Extensive adhesion to diaphragm Extensive adhesion to adjacent organ other than diaphragm Impairment of adjacent major hepatic vein Extensive adhesion to intrathoracic space Operation time, min* Blood loss, mL* $1,000 Usage of RBC transfusion, n Usage of fresh frozen plasma transfusion, n No residual cancer, n R0/R1/R2 Anatomic resection, n Deviation from preoperative planning, n Extended procedure Less-invasive procedure No. of tumors* Maximal size of tumors, mm* Tumor status, n Portal invasion Venous invasion Bile duct invasion Intrahepatic metastasis Grade: well/moderate/poor/ unknown
67 (42–81) 42:4 9/34/3
1 0 0 15 3.9 0.7 88.1
(0–15) (0–1) (0–5) (4.8–45) (2.7–4.7) (0.2–1.8) (63–100) 30/15/1 9.95 (2–18,425) 30 (65%)
27 13 4 2
(59%) (28%) (9%) (4%)
28 (61%) 20 (44%) 5 (11%) 2 (4%) 1 (2%) 424 860 19 3 28
(98–810) (100–3,545) (41%) (7%) (61%)
38 (83%) 38/8/0 28 (61%) 38 (83%) 22 16 1 39
(48%) (35%) (1–11) (10–110)
23 (50%) 7 (15%) 2 (4%) 3 (7%) 5/29/8/4 (continued)
Table I. (continued) n = 46 Noncancerous tissue CH/LC, n Pathological stage I/II/III/IVA, n Morbidity, ny Hospital stay, d*
22/24 4/12/20/10 12 (26%) 16 (7–49)
*Median (range). yComplications of Clavien-Dindo grade $3a. Values in parentheses are percentages, unless indicated otherwise. AFP, Alpha-fetoprotein; CH, chronic hepatitis; ICGR15, indocyanine green retention rate at 15 min; LC, liver cirrhosis; PEIT, percutaneous ethanol injection therapy; RBC, red blood cells; RFA, radiofrequency ablation; TACE, transcatheter arterial chemoembolization; TAI, transcatheter arterial infusion.
grossly and was completely extracted. As for the pathologic findings, microscopic portal venous and hepatic venous tumor invasion were observed in 23 patients (50%) and 7 patients (15%), respectively. Moderately differentiated HCC was most common (n = 29, 63%), whereas well-differentiated HCC rarely was observed (n = 5, 11%). Characteristics of the local and multifocal groups. A comparison of the local and multifocal groups (Table II) revealed that an extended operative procedure was required more frequently in the local group (20/30, 67%), compared with the multifocal group (2/16, 13%) (P < .01). The intraoperative blood loss was greater in the local group than in the multifocal group (1,140 [180–3,110] mL and 435 [100–3,545] mL, respectively; P < .01). The local group required a longer duration of operation (443 [235–783] minutes), compared with the multifocal group (311 [98–810] minutes) (P = .057). In the local group, the median maximum size of the tumors was 27 mm (range, 15–35) before RFA and was 43 mm (range, 10–110) before salvage resection (Table II). Twenty-four patients (80%) exhibited an increase in the size of the local recurrence after RFA. Occult tumor thrombi were noted in 9 cases (30%) before the salvage resection, whereas no tumor thrombi were observed prior to RFA. The tumor-occupying region was distributed more widely before salvage resection than before RFA. The number of cases requiring a resection involving a greater number of Couinaud segments was greater before salvage resection than before RFA (Fig 1, D). Twenty patients (67%) actually required operations that were more extensive than the operations that would have been required before RFA. Although most of the patients (28/30, 93%) fulfilled the commonly adopted criteria for
Yamashita et al 467
Surgery Volume 157, Number 3
Table II. Comparison of operative characteristics between local and multifocal groups Local group (n = 30) Operation time, miny Blood loss, mLy Red blood cell transfusion, mLy Fresh frozen plasma transfusion, mLy Anatomic resection Deviation from the presumed type of resection Extended procedure Less-invasive procedure No residual cancer (R0) No. of tumorsy Maximum size of tumors, mmy Tumor status Portal invasion Venous invasion Bile duct invasion Intrahepatic metastasis Grade: well/moderate/poor/unknown Noncancerous tissue CH/LC Morbidityx Biliary leakage Accumulation of pleural or ascitic fluid Mortality Hospital stay, dy
443 1,140 0 800 24 24 20 4 24 1 43
(235–783) (180–3,110) (0–520) (0–5,040) (80) (80) (67) (13) (80) (1–9) (10–110)
17 (57) 6 (20) 2 (7) 2 (7) 4/16/7/3 15/15 8 (27) 3 (10) 5 (17) 0 (0) 16 (11–44)
Multifocal group (n = 16) 311 435 0 0 4 14 2 12 14 1 26
P value*
(98–810) (100–3,545) (0–0) (0–3,200) (25) (88)
.057z
