Remember me
ORIGINAL ARTICLE
Year : 2022 | Volume
: 59
| Issue : 3 | Page : 368-374
Predictors of toxicity after neoadjuvant chemoradiotherapy for locally advanced gall bladder cancer
Anushree Loyal1, Supriya Chopra2, Mahesh Goel3, Shaesta Mehta4, Prachi Patil4, Shraddha Patkar3, Shyam Shrivastava1, Reena Engineer1
1 Department of Radiation Oncology, Tata Memorial Hospital, Mumbai, Maharashtra, India
2 Department of Radiation Oncology, Advanced Centre for Treatment, Research and Education in Cancer, Navi Mumbai, Maharashtra, India
3 Department of Surgical Oncology, Tata Memorial Hospital, Mumbai, Maharashtra, India
4 Department of Digestive Diseases and Clinical Nutrition, Tata Memorial Hospital, Mumbai, Maharashtra, India
Correspondence Address:
Reena Engineer
Department of Radiation Oncology, Tata Memorial Hospital, Mumbai, Maharashtra
India
Source of Support: None, Conflict of Interest: None
Check
1DOI: 10.4103/ijc.IJC_822_19
Background: The present study evaluated the correlation of hepatobiliary toxicity and radiation dose received in patients undergoing neoadjuvant chemoradiotherapy (NACRT) for locally advanced unresectable gall bladder cancers (LAGBC).
Methods: Twenty-six patients with LAGBC, treated with NACRT (55–57 Gy/25 fractions/5 weeks and weekly gemcitabine 300 mg/m2) within a phase II study, were included. Whenever feasible, surgery was performed after NACRT. Acute and late hepatobiliary toxicity was recorded. Treatment scans were retrieved to delineate central porto-hepatobiliary system (CPHBS), resected liver surface, segment IV B and V, and duodenum. The doses received by these structures were recorded and correlated with toxicity.
Results: Of 26 patients, 20 (77%) had partial or complete response and 12 (46%) had R0 resection. At the median follow-up of 38 months, overall survival was 38%. Eight (30%) patients had post-treatment toxicity, of which most common was biliary toxicity (30%). A correlation was observed between the biliary leak and V45Gy CPHBS >50 cm3 (P = 0.070). Higher toxicity was observed in those with metallic stents (P = 0.072).
Conclusion: The incidence of the biliary leak was 46%. CPHBS dose was found to correlate with biliary leaks. Restricting V45Gy CPHBS <50 cm3 and using plastic stent may facilitate a reduction in hepatobiliary toxicity in patients undergoing NACRT and surgery.
Keywords: Locally advanced gall bladder cancer, neoadjuvant chemoradiotherapy, predictors of toxicity
Key Message The present study aims to investigate factors for increased toxicity after neoadjuvant chemoradiotherapy in locally advanced unresectable gall bladder cancers . This will help in better patient selection and minimal treatment morbidity in future.
Gall bladder cancer is the most common biliary tract cancer[1] and most aggressive amongst the gastrointestinal malignancies with the shortest median survival.[2] Surgical resection with or without adjuvant chemotherapy is the standard of care and the only curative treatment for select patients with gall bladder cancer.[3] Dismal outcomes have been observed in patients with locally advanced disease undergoing nonsurgical management with 1-year survival rates of <10%.[4],[5],[6] A study using radiotherapy in the adjuvant setting has demonstrated that gall bladder cancers are radiation responsive.[7] Focusing high precision and high-dose chemoradiation to the gall bladder and draining lymph nodes can help in downsizing locally advanced tumors within a select proportion (up to 71%) and making them amenable for surgical resection.[8] In this study of locally advanced cancers, the 5-year overall survival was 47% in those who underwent R0 surgical resection after neoadjuvant chemoradiotherapy (NACRT). While the survival outcomes following NACRT are encouraging, a prolonged biliary leak was observed in up to 46% of patients undergoing surgical resection after NACRT, which led to substantial morbidity and prolonged hospitalization. Since biliary leaks were a major cause of morbidity, the present study was undertaken to investigate factors impacting acute and late biliary toxicity (or morbidity) in patients undergoing neoadjuvant chemoradiation and surgery for locally advanced gall bladder cancer.
MethodsThe present study aimed to systematically investigate the impact of patient and treatment-related factors (like the presence of plastic or metal stent, radiation dose to the central porto-hepatobiliary system [CPHBS], radiation dose to the resected liver surface) on acute and chronic biliary toxicity in patients undergoing NACRT and surgery for locally advanced gall bladder cancer. The present study is a part of prospective study which was approved by the Institutional Review Board and Human Ethics Committee of Tata Memorial Hospital, and written informed consent was taken from all patients. The Data and Safety Monitoring Committee reviewed the study with regard to the safety of the protocol. The date of approval is 29 October 2008, CTRI/2016/08/7129.
The treatment records and radiation planning details of patients treated within the prospective registered phase II study were retrieved (NCT01118897). Within this prospective study, patients with locally advanced gall bladder cancer, as per the TNM classification of the American Joint Committee on Cancer 6th edition, with stage III disease (with deep liver infiltrations or portal nodes, or both), were treated with concurrent chemoradiation using helical tomotherapy (dose of 57 Gy/25 fractions to the gross tumor and 45 Gy/25 fractions to the surrounding nodes over 5 weeks) with weekly injectable gemcitabine (300 mg/m2). N1 nodes around the cystic duct, pericholedochal, or hilar lymph nodes (i.e., in the hepatoduodenal ligament) and N2 nodes such as peripancreatic, periduodenal, and periportal nodes were considered locoregional and included in the study. Patients with partial or complete response to treatment, as assessed by positron emission tomography (PET) or contrast-enhanced computed tomography (CECT) 6 weeks after radiotherapy, were considered for surgery. The surgical procedure planned was en bloc resection, including cholecystectomy with a liver wedge excision and periportal lymph node clearance. The use of adjuvant chemotherapy was left to the decision of the treating physician.
The incidence of biliary toxicity (or leaks) following surgical resection was obtained from the prospective trial database. Common terminology criteria for adverse events v4.03 were used for toxicity grading. Patients were categorized as those with or without biliary toxicity for statistical analysis. Patient-related factors like age, presence of morbidity (diabetes, hypertension), and hypoalbuminemia (yes or no) were categorized for toxicity analysis. Also, the type of biliary stent placed prior to surgery (plastic or metallic) if any was identified. The radiation plans used for treatment delivery were retrieved, and the central porto-hepatobiliary tract was delineated on day 1 megavoltage CT (MVCT) on tomotherapy workstation (Hi-Art Version 4.2.1, Tomotherapy Inc., USA) with an aim to record the dose received by the CPHBS. The main portal vein was identified and delineated on serial CT images. The main portal vein structure was expanded using a brush, with a 0.5 cm radius, to delineate the CPHBS [Figure 1]. As surgery for locally advanced gall bladder cancer often involved wedge resection of segment IVb and V; therefore, the excised liver segment was delineated using information from the surgical and histopathology report [Figure 2]. The width and height of the delineated liver segment corresponded to the gross excised liver wedge adjacent to the original tumor location. Segment IVb and V of the liver were also delineated [Figure 3]. Duodenum was contoured according to the radiation therapy oncology group upper abdominal normal structure contouring guidelines.[9]
Figure 1: Portal vein and central porto-hepatobiliary system (CPHBS) delineation, giving a margin to portal vein using 0.5 cm brush
Figure 2: Axial CT image showing resected liver surface (adjacent to GTV gall bladder). CT: Computed tomography; GTV: Gross tumor volume
Figure 3: All the contours in single axial CT slice (GTV—red, excised liver surface—green, segment IVb—pink, segment V—violet, CPHBS—purple). GTV: Gross tumor volume, CPHBS: Central porto-hepatobiliary system, CT: Computed tomographyAfter delineation, the doses received by CPHBS and excised liver (including the resected margin) were determined on tomotherapy adaptive workstation. The volume of CPHBS receiving a dose of 5–60 Gy was recorded at every 5 Gy interval (i.e., V5, V10, V15 … V60 Gy). SPSS version 21.0 software program (SPSS Inc., Chicago, IL, USA) was used for statistical analysis. A receiver operating characteristic curve identified CPHBS and liver segment dose-volume thresholds that predicted for grade 3 or higher toxicity with the highest specificity. All data were dichotomized across these identified cutoff values for univariate analysis. The impact of patient and treatment-related factors on biliary toxicity was assessed using univariate and multivariate analysis.
ResultsOf the 28 patients accrued in the primary study, treatment planning scans of 26 patients could be retrieved, who formed the study cohort for toxicity analysis. Treatment planning scans could be retrieved for 26/28 patients who formed the study cohort. The median age was 53.5 years (31–69 years). Of the 26 patients, 8 (31%) patients had T2 disease, and 18 (69%) patients had T3 disease at presentation. Nine patients (35%) were node negative at presentation. N1 disease was seen in 6 (23%) patients, while 11 (42%) patients had N2 disease, which precluded R0 resection. The mean tumor size was 3.5 cm (1.5–7 cm). Overall, 14 patients (54%) had deep liver infiltration. Four patients had hilar invasion. Three patients had colon involvement, and three patients had duodenal wall infiltration on imaging. These patients with bowel involvement on imaging underwent endoscopic examination and were included in the study only after ruling out any mucosal involvement, to avoid the risk of fistula postradiotherapy. Four (15%) patients had stent placed pretreatment, indicative of obstruction at presentation. These four patients had extrahepatic biliary invasion involving the hilum. Two patients had type I block, one had type II block, and another had type IIIA block, as classified by the Bismuth classification. All four patients had self-expandable metallic stent (SEMS). All the patients accrued in the study, except three, completed radiotherapy. Amongst the three patients who could not complete radiotherapy, one patient developed cholangitis, one developed colonic perforation, and one had disease progression (liver metastasis) on radiotherapy.
Of the 26 patients, 9 (35%) patients had complete response postchemoradiotherapy, while 10 (38%) patients had a partial response and 1 (4%) patient had stable disease. A total of six (23%) patients had disease progression during or within 6–8 weeks of treatment completion as assessed by postchemoradiotherapy imaging (PET-CT or CECT). Overall, 20/26 (77%) patients had a response to treatment and were considered for surgical exploration, of which 2/20 patients refused surgery. A total of 18 (69%) patients were considered for surgical resection, of which surgery was abandoned intraoperatively in 5 patients due to peritoneal or liver metastasis. Overall, 12 (46%) patients had R0 and 1 (4%) had R2 resection. Nine patients received gemcitabine-based adjuvant chemotherapy. Patients who had disease progression or no response after chemoradiotherapy were considered for palliative chemotherapy.
Toxicity analysis showed toxicity in 8/26 (30%) patients post-treatment [Table 1]. The most common toxicity was biliary toxicity in eight patients. One patient had cholangitis during RT and biliary stricture (grade 3) developed in one patient, 5 months postradiotherapy. Six patients (23%) had postoperative bile leak (duration 15 days to 1 year). Grade 3 bile leak was seen in two patients requiring intervention while the remaining four patients had grade 1 bile leak. Of these, bilio-cutaneous fistula developed in two patients. The site of postoperative bile leak was not categorized as peripheral or central bile leak in the study protocol. Two patients with SEMS having biliary leak, the texture of the bile duct was paper-thin intra-operative, of which one required endoscopic retrograde cholangiopancreatography and stent placement, and the other had to undergo percutaneous transhepatic biliary drainage and stenting. The remaining four patients developing biliary leak who did not have SEMS were managed conservatively. Magnetic resonance cholangiopancreatography carried out in these patients could not localize the site of the leak to the biliary tree; therefore, the leaks were assumed to be from the liver wedge area, all of which stopped with conservative management.
Table 1: Number of patients and post-treatment toxicity (gall bladder adenocarcinoma, n = 26)Toxicity was correlated with patient and tumor characteristics. The median age of the patients with and without biliary toxicity was 50 ( range: 31-55) and 57.5 (range: 40-69) years, respectively (P = 0.030), suggesting more toxicities in the younger age group. The dose-volume correlation for biliary leaks was restricted to patients who completed neoadjuvant chemoradiation and surgery. Out of 13 patients who underwent surgery, 2 had stent. Both these patients had postoperative bile leak. However, in the remaining 11 patients, only 4 patients developed postoperative bile leak (100% versus 36%, P = 0.192). The presence of hepatic infiltration (P = 0.237) and hilar invasion (P = 0.192) did not significantly influence the rate of bile leak. No correlation was seen between nodal positivity and the rate of bile leak (P = 1). Patients with N1 disease (periportal node) had a higher biliary leak rate (3 out of 4); however, there was no statistically significant correlation (P = 0.254).
No significant correlation was found between the dose received by the various volumes of the CPHBS and the biliary leak. However, a weak positive correlation was found between the rate of biliary leak and V45 Gy > 50 cm3 (P = 0.070) and higher volume bins of V50 Gy and V55 Gy [Table 2]. A trend toward increased toxicity, although not statistically significant, was seen with the CPHBS D1cc > 58 Gy (EQD2(3) 61.7 Gy) (P = 0.103) and D5cc > 55 Gy (EQD2(3) 57.2 Gy) (P = 0.286). The rate of postoperative bile leak was not associated with the median doses received by the resected liver surface (P = 0.192), segments IVb (P = 0.266), and V (P = 0.592). Two patients had duodenal toxicity in the form of duodenal stenosis or gastric outlet obstruction postradiotherapy, both requiring stenting for the relief of obstructive features. V55 (EQD2 (3) 57.2 Gy) for patients without duodenal toxicity was 6.2 cm3, while it was 7.9 cm3 for patients with toxicity (P = 0.772). Duodenum D1cc was also higher in the cohort with duodenal toxicity (D1cc = 58.6 Gy) as compared to the cohort without toxicity (D1cc = 54.3 Gy) though it was not found to be statistically significant (P = 0.248). Similarly, D5cc was high in the toxicity group, but this did not reach statistical significance (P = 0.500) [Table 3].
Table 2: Dosimetric parameters of patients who underwent surgery and correlation with biliary toxicity (gall bladder adenocarcinoma, n = 13)
Table 3: Dosimetric parameters and correlation with duodenal toxicity (gall bladder adenocarcinoma, n = 26)At a median follow-up of 38 months (range: 2-56 months, June 2009–December 2014), out of 26 patients, 17 patients had disease progression. Overall, 7 patients are alive without any evidence of disease, 3 patients are alive with disease, 14 patients have died of disease, and 2 patients died of other causes.
The median overall survival for the patients included in the study is 16 months (2–53 months). When compared between the patients who had surgery post-NACRT and those who did not have, a statistically significant difference was seen in the two subgroups (77% versus 0%) (no patient survived) (P = 0.000) with improved rates of survival in patients in the surgery arm [Figure 4]. No impact of nodal status was observed on overall survival (P = 0.878). The overall survival was 38% at the median follow-up of 38 months. For patients who underwent surgery (R0), the survival was 75% after a median follow-up of 38 months.
Figure 4: Surgery and overall survival (gall bladder adenocarcinoma, n = 26). Arm; number of patients: surgery 13; and no surgery 13 Discussion There are no robust data on the use of NACRT for locally advanced gall bladder cancer. A study from our institution has shown that locally advanced unresectable cancers may benefit from neoadjuvant chemoradiation to facilitate a curative resection with a good survival but with increased toxicity.[8] The present study was undertaken to investigate factors impacting acute and late toxicity (or morbidity) in patients undergoing neoadjuvant chemoradiation and surgery for locally advanced gall bladder cancer. In the present study, we observed a clinically significant rate (77%) of surgical conversion in patients presenting with unresectable disease. Of the 20/26 patients who became resectable, surgery was abandoned in 5 patients due to intraoperative finding of disease progression in the peritoneum, suggesting that eventually up to 50% of patients could be successfully downstaged and operated. This was, however, at the cost of significant biliary toxicity (biliary leaks, strictures, and cholangitis) seen in up to one-third of the patients. A few patients had more than one toxicity, which included duodenal stenosis and gastric outlet obstruction.
The rate of postoperative or postradiotherapy biliary toxicity has been reported to range from 3% to 23%.[10] Bartlett et al.[10] reported 26% morbidity in patients undergoing surgery for gall bladder cancer, with the majority being in patients who had bile duct excision. In another study by Dixon et al.,[11] 49% morbidity was reported with extended resections; however, only 29% were major, requiring intervention. The rate of biliary leak was reported to be 2% only. However, surgical series have reported overall operative morbidity of 10–13% in patients who underwent radical cholecystectomy without extended resection[12],[13] and the rate of biliary leak as 2–5%.[11],[14] However, in the present study, the rate of biliary toxicity was 30% (biliary leak 23%), higher than that reported in any of the published series or that observed in patients undergoing upfront surgery (6%).
In the present study, younger age was associated with higher toxicity (P = 0.030) and could be attributed to more advanced cases with obstructive features observed in the younger age group in the present study. Various studies have reported advanced age to be associated with bile leakage, which contrasts with what has been observed in this study.[15],[16] Overall, three out of four patients with biliary stent had biliary toxicity. This could be attributed to the use of a metallic stent that causes more dose deposition in the adjacent structures. Studies in esophageal cancer suggest that the presence of metallic stent increased the mucosal radiation dose by 5–10% at a depth of 5 mm,[17],[18] due to an increase in absorption of scattered dose around the metallic stent. Though statistically nonsignificant (P = 0.192), bile leak was observed in increased frequency in patients with preradiotherapy stent. Also, at the time of surgery, the bile duct consistency was found to be paper-thin, which could be a consequence of increased dose to the biliary duct. So, these results suggest that extra care should be taken for dose evaluation in patients with a metallic stent.
Patients with hilar invasion or invasion of the porta hepatis had a trend toward increased biliary toxicity (P = 0.072). This variable has not been taken into consideration in other studies. Thus, hilar invasion is of uncertain value concerning biliary toxicity. However, the evidence of hilar invasion on imaging suggests a possibility of bile duct being present in the region of high dose that received a simultaneous integrated boost. Such high doses to the biliary tree could be a plausible explanation for increased biliary toxicity in these patients.
There was no significant correlation of biliary toxicity with hepatic infiltration (P = 0.473). No study has reported any correlation with regard to these two variables. However, studies have shown bile leak to be associated with a cut surface area of ≥57.5 cm2.[15],[16],[19] The patients with hepatic infiltration at the time of surgery might end up with excision of a large surface area of the liver wedge. When correlated with bile leak in patients who had undergone surgery, this factor remained insignificant (P = 0.237). Hence, no inference can be drawn.
The doses received by CPHBS were correlated with the rates of biliary toxicity. No statistically significant difference was obtained in patients with or without biliary toxicity. However, drawing definite conclusions from the above results would possibly result in bias due to the small sample size. It was observed that larger volumes irradiated to high dose (V55 Gy > 53 cm3) had a trend toward increased risk of biliary toxicity (P = 0.086). In subgroup analysis in patients with surgery post-CTRT, the CPHBS V45 Gy > 50 cm3 had a weak positive correlation (P = 0.070) with biliary toxicity. So these results, though not statistically significant, suggest that irradiating large volumes of the biliary tree with high doses could be a predisposing factor to biliary toxicity. Also, from the above data, it can be hypothesized that in patients who undergo surgery, doses higher than 45 Gy predispose to biliary toxicity. In contrast, patients who do not undergo surgery tolerate doses up to 50–55 Gy well. These findings prompt the need for future studies in a large number of patients.
On correlating the biliary toxicity with CPHBS dose, a weak positive correlation with D1cc > 58 Gy was observed (P = 0.103). No study could be found in the literature, which suggests a correlation of biliary toxicity with radiotherapy dose after fractionated radiotherapy; however, a few studies have reported higher rates of biliary toxicity post-stereotactic body radiotherapy.[20],[21]
Two patients had duodenal toxicity. The V55 Gy for patients without duodenal toxicity was 6.2 cm3, while for patients with duodenal toxicity, it was 7.9 cm3 (P = 0.772). These findings were not significant and did not suggest a correlation of duodenal toxicity with the doses received.
The median overall survival at a median follow-up of 38 months (June 2009–December 2014) was 16 months (2–53 months). The overall survival was significantly better in the patients who had undergone surgery as compared to those who had no surgery or in those who did not respond well to chemoradiotherapy (P < 0.0001). This is keeping with the studies in the literature that have reported complete or an R0 resection an important prognostic factor[22] and its association with improved survival. Nodal status had no impact on overall survival in the present study (P = 0.878).
The main limitation of this prospective study is its small sample size. Also, the additional structures were contoured retrospectively on MVCT on the adaptive planning station. Since the dose calculation algorithm used IVDT (image value-to-density table) table for MVCT image, as against the kilovoltage CT IVDT table used on planning CT, there may be dose variation of around 2–3% in the calculated and the actual dose received, though it lies within acceptable limits.[23] Furthermore, there may be some dose variation due to target volume displacement as a result of internal organ motion, which may have been unaccounted. Also, no recommended guidelines exist for the delineation of the biliary tree. We contoured the biliary tree based on its anatomical relationship with the portal vein, and this remains to be validated.
To summarize, a marked improvement in the outcome of patients with gall bladder cancer was achieved in patients who underwent surgery, especially R0 surgical resection. After NACRT, 50% of the cases of locally advanced gall bladder cancer became resectable. However, it was associated with increased postoperative biliary toxicity. Analysis of dosimetric data of normal tissue structures revealed no significant correlation with toxicity or adverse events. However, it was evident that doses to the biliary tree more than 45 Gy in patients undergoing surgery and more than 55 Gy in patients undergoing CTRT alone were associated with more toxicity. Also, pretreatment biliary metallic stent was associated with increased toxicity. So, to conclude, further systematic research incorporating all set uncertainties and risk factors like stent in situ, in a comprehensive manner on a larger cohort of patients, is warranted.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References
Comments (0)