Salvage hepatectomy for local recurrence after particle therapy using proton and carbon ion beams for liver cancer

Abstract Aim With the increased use of particle therapy for liver cancer, local recurrence after particle therapy increased. Salvage hepatectomy is an acceptable treatment option for local recurrence following particle therapy; however, its safety and effectiveness remain unclear. Therefore, this multi‐center study aimed to verify the feasibility and efficacy of salvage hepatectomy and assess clinical issues associated with its application. Methods We retrospectively assessed the perioperative outcomes, prognosis, and pathological characteristics of 15 patients who underwent salvage hepatectomy for local recurrence after particle therapy between 2006 and 2019. Results Hepatocellular carcinoma and metastatic liver tumors were noted in eight and seven patients, respectively. The mean total dose and number of fractions were 66.5 Gy and 12, respectively, and the mean interval between particle therapy and surgery was 30.1 months. Major hepatectomy was performed in seven cases. Moreover, the mortality rate was 0%, and surgical complications of Clavien‐Dindo grade IIIa or higher were observed in four cases (27%)—two bile leakages, one pleural effusion, and one refractory skin fistula. The median overall survival time and 5‐year overall survival rate after salvage hepatectomy were 29.9 months and 43.1%, respectively. Histological examination of the irradiated liver tissue surrounding the tumor showed sinusoidal dilatation, loss of hepatocyte, and fibrosis in most cases. Conclusion Salvage hepatectomy after particle therapy is a feasible therapy; however, the risk of refractory complications associated with particle therapy is relatively high. Therefore, the first‐line treatment for resectable liver cancer should be carefully determined considering second‐line treatment after local recurrence.


| Patients
A multi-center retrospective study was conducted. Clinical and pathological data were collected and analyzed.
This study was approved by the Institutional Review Board of each institution.

| Data collection
Clinical characteristics, such as age, sex, liver function, and tumor factor and type, were collected. Additionally, PT information, such as beam type, intensity, fraction, radiation planning, and interval time from irradiation to surgical operation, were also obtained.
Moreover, postoperative outcomes, such as morbidity, mortality, survival status, and recurrence status, were obtained and analyzed.

| Statistical analyses
The follow-up time was calculated from the day of salvage hepatectomy. The probabilities of overall survival (OS) and recurrence-free survival (RFS) were estimated using the Kaplan-Meier method and compared using the log-rank test. The correlation between the irradiation dose and the resection line and postoperative complications was estimated using the chi-square (χ 2 ) test. All statistical analyses were performed using the SPSS statistics software version 22.0 (IBM Corp., Armonk, NY, USA).

| Pathological examination
Resected specimens were examined for pathological changes of the recurrent tumor and the adjacent non-cancerous liver tissue by applying PT. Staining methods, such as hematoxylin and eosin (H&E) and Masson, were employed.

| Clinical characteristics of the study participants
The details of patients who underwent salvage hepatectomy are provided in Table 1   adenocarcinomas. The median size and number of primary tumors before PT were 3.5 (range: 1.0-9.7) cm and 1.1 (range: 1-2), respectively.
Particle therapy was conducted as the primary treatment for patients who wanted to undergo this procedure in 13 cases and in two cases of unresectable HCC. The former 13 patients had resectable tumors; however, they chose to undergo PT as a less invasive therapy. The latter two cases were of a patient with HCC with portal venous tumor thrombus and another with HCC with lung metastasis.
Since the portal venous tumor thrombus was shrunk by RT and R0 resection could be performed for recurrent tumor, and lung metastasis was already resected after PT without recurrence, we decided to perform salvage hepatectomy. PT was performed with curative intent using protons or carbon ions. The mean total dose and num-

| Outcome of salvage hepatectomy
The surgical outcomes of salvage hepatectomy are summarized in Table 2. Major hepatectomy (more than two sections) was performed in seven patients. The mean operation time and blood loss were 396 (range: 143-654) minutes and 402 (range: 66-1000) mL, respectively.
No postoperative death was recorded.
Surgical complications occurred in six cases (40%). Clavien-Dindo grade IIIa or higher was observed in four cases (27%), consisting of two HCC and two metastatic liver tumor cases: two, one, and one cases of bile leakage, pleural effusion, and refractory skin fistula, required abdominal wall reconstruction using rectal muscle flap. The mean postoperative hospitalization duration of these four patients with Clavien-Dindo grade IIIa or higher was 59 (range: 13-126) days; two patients with refractory bile leakage, after partial resection, required an extended hospital stay of 126 and 82 days.

| Prognosis after salvage hepatectomy
The prognosis after salvage hepatectomy is presented in Figure 1.
The median OS and RFS of all patients in this study were 29.9 and 19.9 months, respectively; the 5-year OS and RFS rates were 43.3% and 30%, respectively ( Figure 1A,B). In patients with HCC, the median OS and RFS were 23.9 and 19.9 months, and in patients with liver metastasis, the median OS and RFS were 29.9 and 11.0 months, respectively. There were no significant differences in the OS and RFS between HCC and liver metastasis (P = .85 and .86; Figure 1C,D).

| Macroscopic and microscopic findings
A typical case of salvage hepatectomy is presented with macroscopic and microscopic findings in Figure 2. A 66-year-old male patient initially underwent PT with proton beam for liver metastasis from rectal cancer (Case 7 in Table 2). The radiation intensity was 66 Gy (RBE) in 10 fractions (Figure 2A). Following 24 months of PT, tumor regrowth was detected on computed tomography ( Figure 2B), and the patient was referred to our hospital. Intraoperative findings revealed a strong adhesion between the liver and the diaphragm ( Figure 2C); consequently, central bisegmentectomy was performed ( Figure 2D).
In recurrent tumors, moderately differentiated tubular adenocarcinoma ( Figure 2E) along with the surrounding necrotic tissue ( Figure 2F) were microscopically observed. Irradiated non-cancerous liver tissue displayed unique pathological changes, such as sinusoidal dilatation ( Figure 2G), loss of hepatocyte ( Figure 2H), and fibrosis of the portal region ( Figure 2I). In contrast, the non-irradiated liver tissue in the lateral segment displayed normal findings ( Figure 2J).
These pathological changes were observed in the majority of the cases.

| Refractory complications
We observed two cases with refractory complications, requiring elongated hospital stay. One case displayed bile leakage (Case 4 in Table 2). A 77-year-old male patient was diagnosed with solitary liver metastasis at segment 6, following 4 years of pancreatoduodenectomy for pancreatic head cancer. The patient opted to undergo PT with carbon ion beams with 60 Gy (RBE) in eight fractions ( Figure 3A). After 7 months, tumor regrowth was detected by computed tomography ( Figure 3B); he was subsequently referred to our hospital. After the evaluation of the tumor status and liver function, we determined that the regrowth tumor was resectable; thus, partial resection in segment 6 was performed. Postoperatively, refractory bile leakage, subsequent local infection, and intra-abdominal abscess formation were observed; drainage tube placement was required for a prolonged period ( Figure 3C). Finally, the patient was discharged on postoperative day 126. A retrospective review of the dose distribution in radiation planning indicated the resection line of the liver parenchyma overlapped with the 80% dose irradiated area ( Figure 3A,B).
The other postoperative complication observed was refractory skin fistula (Case 14 in Table 2). A 54-year-old male patient was diagnosed with HCC associated with chronic hepatitis B. The tumor was solitary, 3.6 cm in size, located in segment 3, and was resectable; however, as per the patient's request, PT with 66 Gy (RBE) of proton beam in 10 fractions was performed ( Figure 4A).
After 10 months, local recurrence was detected; then, salvage left hepatectomy was performed with a reverse T-shaped skin incision. Following surgery, wound dehiscence and skin fistula occurred on the midline incision, and it was refractory ( Figure 4B).  Figure 4A); skin discoloration by irradiation was observed ( Figure 4B).

| Association between irradiation dose of resection line and refractory complication
We investigated dose distribution in radiation planning in all cases; the association of the effect of increased irradiation with the incision line of the liver or the skin with refractory complications was analyzed. When the cutoff value was set to a maximum dose of 70%, an increased amount was significantly associated with Clavien-Dindo grade IIIa or higher complications; four (66%) out of six patients and none out of nine patients who received >70% and <70%, respectively, presented grade IIIa complications or higher (P = .01).

| D ISCUSS I ON
In the present study, we reported the outcomes of salvage hepatectomy following PT for liver cancer with a certain number of cases.
Salvage hepatectomy appears to be feasible; however, there are specific aspects to be considered for surgical indications and procedures to prevent refractory complications.
After PT, surgical resection becomes more challenging because inflammatory changes occur at irradiated areas. As shown in of the detachment of the liver parenchyma from the diaphragm and the major vessels was required. In this study, major hepatectomy was performed in about half of the cases; consequently, the operation time and blood loss were acceptable.
We experienced a relatively elevated rate of Clavien-Dindo grade IIIa complications or higher: two bile leakages, one skin fistula, and one pleural effusion. Bile leakage cases were both of partial resection, in which the parenchymal dissection line was on a high-dose irradiated area. The resultant skin fistula was a case of an open left hepatectomy with a midline incision of the skin; a high-dose proton beam irradiation was performed. In this case, the irradiated skin was discolored at the time of hepatectomy. As a result of refractory complications, it was speculated that wound healing did not occur effectively due to an excessive dose of particle beam irradiation. In this study, the cutting of the liver or skin in a high-dose irradiated area was significantly associated with severe complications. To pre- Another concern was the safe radiation dose of the particle beam for salvage hepatectomy. In the present study, 70% of the maximum dose is significantly associated with severe complications. However, there were no sufficient data regarding the safe radiation dose for surgical dissection. Therefore, further investigation is warranted to explore this issue.
There were few reports about the pathological changes in the liver following PT, whereas conventional radiotherapy was outlined. 11 In this study, the irradiated non-cancerous liver tissue surrounding the tumor displayed sinusoidal dilatation, loss of hepatocyte, and fibrosis in the majority of the cases. As for the pathogenesis of liver damage after irradiation, a non-human primate model revealed conventional radiation-induced veno-occlusive changes, resulting in sinusoidal congestion and fibrosis. 12 Interestingly, in the present study, liver function was well-preserved after PT in the majority of the patients; hence, it was deemed feasible for major salvage hepatectomy in half of the cases. This observation may depend on the high-dose concentration of PT for liver tumor. Normal pathological findings in the non-irradiated liver tissue outside the irradiated area provided suitable evidence ( Figure 2J).
With regard to prognosis, the 5-year OS and RFS rates in all cases were 43.3% and 30%, respectively; there was no significant difference between HCC and liver metastasis. The 5-year OS rate of salvage hepatectomy for recurrent HCC following locoregional therapy, such as radiofrequency ablation and/or transarterial chemoembolization, was 38%-69%. 13,14 Although the number of patients in the present study was limited, we found that salvage hepatectomy is a potential rescue therapy to prolong survival following local recurrence of HCC and metastatic liver tumor after PT.
According to the guidelines of the American Association for the Study of Liver Diseases and the Japan Society of Hepatology, the methods for HCC treatment with curative intent include surgical resection, radiofrequency ablation, and liver transplantation. 15,16 Radiation therapy, including PT, has not been standardized in these therapeutic algorithms yet. However, a positive local control rate for HCC has been reported. [3][4][5][6] This includes advanced cases with portal venous 17 or inferior vena cava tumor thrombus. 4 Moreover, a non-randomized controlled study comparing proton beam therapy to surgical resection for resectable HCC is ongoing in Japan. 18 For colorectal liver metastasis, surgical resection remains the standard treatment 19 ; however, the local control rate of particle therapy for liver metastasis has been reported to be 61%-66%, 20,21 and several studies displayed an improved local control rate by dose escalation manner with carbon ion and proton beam. 22 The limitations of the present study were its retrospective nature and the small sample size. Currently, the number of salvage hepatectomies conducted after PT is limited and only a few case reports are referable. 8,9 In 14 years, salvage hepatectomy was performed in only 17 out of 110 patients who were treated with PT in Hyogo Ion Beam Center and experienced local recurrence. There are some reasons for this small number: difficult pattern of local recurrence, liver dysfunction after PT, and less knowledge concerning salvage hepatectomy. However, the number of patients undergoing PT for liver cancer has increased in the last two decades 3 because of improved techniques, higher numbers of facilities, and expanded use for resectable cases and metastatic liver tumor. These circumstances add weight to salvage hepatectomy in the treatment of recurrent liver tumor.
The outcomes obtained from this study can be a guide for surgeons to perform proper treatment selection and perioperative management of local recurrence after PT. We believe that this study may contribute to spreading the knowledge about salvage hepatectomy.
In conclusion, salvage hepatectomy for local recurrence after PT is feasible and effective; however, it generates an increased risk of refractory complications associated with previous irradiation.
Therefore, the first-line therapy for resectable liver tumor should be carefully determined, considering the possibility and risk of secondline therapy.

D I SCLOS U R E
Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.