Modified cisplatin/interferon α-2b/doxorubicin/5-fluorouracil (PIAF) chemotherapy in patients with no hepatitis or cirrhosis is associated with improved response rate, resectability, and survival of initially unresectable hepatocellular carcinoma
Ahmed O. Kaseb MD,
Department of Gastrointestinal Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas
Department of Surgical Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas
Corresponding author: Jean-Nicolas Vauthey, MD, Department of Surgical Oncology, Unit 1484, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030; Fax: (713) 745-1921; email@example.com.
The purpose of this study was to evaluate the factors associated with response rate, resectability, and survival after cisplatin/interferon α-2b/doxorubicin/5-fluorouracil (PIAF) combination therapy in patients with initially unresectable hepatocellular carcinoma.
The study included 2 groups of patients treated with conventional high-dose PIAF (n = 84) between 1994 and 2003 and those without hepatitis or cirrhosis treated with modified PIAF (n = 33) between 2003 and 2012. Tolerance of chemotherapy, best radiographic response, rate of conversion to curative surgery, and overall survival were analyzed and compared between the 2 groups, and multivariate and logistic regression analyses were applied to identify predictors of response and survival.
The modified PIAF group had a higher median number of PIAF cycles (4 versus 2, P = .049), higher objective response rate (36% versus 15%, P = .013), higher rate of conversion to curative surgery (33% versus 10%, P = .004), and longer median overall survival (21.3 versus 10.6 months, P = .002). Multivariate analyses confirmed that positive hepatitis B serology (hazard ratio [HR] = 1.68; 95% confidence interval [CI] = 1.08-2.59) and Eastern Cooperative Oncology Group performance status ≥ 2 (HR = 1.75; 95% CI = 1.04-2.93) were associated with worse survival whereas curative surgical resection after PIAF treatment (HR = 0.15; 95% CI = 0.07-0.35) was associated with improved survival.
Hepatocellular carcinoma (HCC) is potentially curable by surgical resection,[1, 2] local ablation,[3, 4] or liver transplantation.[5, 6] However, most patients with HCC present with unresectable disease accompanied by background liver disease that has an independent effect on patients' survival. Therefore, effective neoadjuvant therapeutic options are limited for these patients.
Conventionally, HCC has been considered a chemotherapy-resistant tumor, and numerous clinical trials of a wide variety of chemotherapeutic and hormonal agents have failed to show satisfactory results. However, multiple expert panels concluded that the underlying liver disease is a major confounding factor that independently affects HCC treatment outcome. This fact is very pertinent to failed chemotherapy trials in HCC, because the majority of patients did not tolerate full doses and regular schedules of chemotherapy, and therefore, the poor outcome and short survival were related, in part, to deterioration of the underlying liver disease. Notably, the PIAF regimen, consisting of cisplatin, interferon α-2b, doxorubicin, and 5-fluorouracil (5-FU), produced a relatively high objective response rate in the initial phase 2 study of this regimen for patients with unresectable HCC. Although a phase 3 clinical trial comparing PIAF with doxorubicin did not show an overall survival benefit of PIAF, the objective response rate of the regimen was confirmed. However, the potential neoadjuvant role of PIAF in unresectable HCC remained an open question because the populations in previous studies had high prevalences (> 80%) of cirrhosis and hepatitis B,[8-11] both of which have been reported to negatively affect the course of intensive chemotherapy and ultimately patients' surgical candidacy.[10, 12] Since 1994, our group has prospectively followed up all the patients who have received PIAF treatment at our institution for initially unresectable HCC. From January 1994 through May 2003, PIAF was administered in the context of an expanded institutional phase 2 trial. In June 2003, on the basis of the initial results of our phase 2 trial and that of another group, the patient selection criteria for PIAF and the PIAF treatment protocol were modified.
The objectives of the current study were to compare the response rate, resectability, and survival between the patients treated with the original PIAF regimen and the optimally selected patients treated with the modified PIAF regimen and to explore the factors associated with survival, response to the PIAF regimen, and improved resection rate.
MATERIALS AND METHODS
This study was approved by the Institutional Review Board at The University of Texas MD Anderson Cancer Center, Houston, Texas, and was conducted in accordance with the principles of the Declaration of Helsinki. The study population consisted of 2 groups of patients: the conventional PIAF group (n = 84) and the modified PIAF group (n = 33) (Table 1). The conventional PIAF group was a prospective cohort in an extended single-arm institutional phase 2 trial assessing the efficacy of the PIAF regimen and conducted between January 1994 and June 2003. The modified PIAF group was a prospective registry cohort treated with the modified PIAF regimen between July 2003 and June 2012. We performed a retrospective analysis of both cohorts.
HCC was diagnosed by histologic examination of tumor tissue. Treatment options for every patient were discussed during a multidisciplinary Liver Tumor Study Group, and resectability was assessed by experienced hepatobiliary surgeons on the basis of imaging findings (eg, tumor size, tumor location, tumor distribution, invasion, and presence of extrahepatic disease), function of the underlying liver, presence of medical comorbidities, and patient performance status.
Indications for PIAF Treatment
General indications for PIAF treatment were as follows: unresectable advanced HCC, life expectancy more than 12 weeks, platelet count of 100,000/mm3 or greater, absolute granulocyte count of 1500/mm3 or greater, serum bilirubin level of 2 mg/dL or less, serum albumin level of 3 g/dL or greater, serum creatinine level of 1.5 mg/dL or less, normal cardiac function, and Eastern Cooperative Oncology Group (ECOG) performance status score 2 or less. Since June 2003, on the basis of results of the phase 2 trial, we adopted the following as relative contraindications to PIAF treatment at our institution: positive hepatitis B serology (ie, presence of hepatitis B surface antigen or anti-hepatitis B core antibody) with no ongoing antiviral therapy, positive hepatitis C serology (ie, presence of anti-hepatitis C antibody), histologic or clinical evidence of established cirrhosis, and ECOG performance status of 2 or greater.
The conventional PIAF regimen consisted of cisplatin at 80 mg/m2 on day 1, interferon α-2b at 5 MU/m2 on days 1 through 4, doxorubicin at 40 mg/m2 on day 1, and 5-FU at 500 mg/m2 over 24 hours on days 1 through 4. The modified PIAF regimen consisted of cisplatin at 20 mg/m2 on days 1 through 4, interferon α-2b at 4 MU/m2 on days 1 through 4, doxorubicin at 40 mg/m2 on day 1, and 5-FU at 400 mg/m2 as a bolus infusion on days 1 through 4. Each treatment was repeated every 4 weeks for up to 6 cycles. Several patients with good response and tolerance received up to 2 more cycles of PIAF after completion of the 6 cycles of PIAF that were part of the planned protocol treatment. In patients who developed grade 3 or higher drug-related adverse events, treatment was withheld until the event resolved to grade 1 or less, and doses were reduced by 25% in the next cycle. Patients who completed scheduled PIAF treatment but still had unresectable viable tumors received additional therapies including transarterial chemoembolization (TACE) and/or other systemic therapy, in consideration of their oncologic status and physical conditions.
Follow-Up and Evaluation of Response
Liver function tests, electrolytes, and serum alpha-fetoprotein level were examined every 4 weeks during PIAF treatment. A baseline computed tomography (CT) scan was obtained before day 1 of the first PIAF cycle, and the disease status was assessed by enhanced CT every 8 weeks (= 2 cycles) during PIAF treatment. Response to chemotherapy determined by an investigator was confirmed by radiologists and described according to the Response Evaluation Criteria in Solid Tumors (RECIST). Objective response rate was defined as the percentage of patients who had a complete response or partial response during PIAF treatment. Toxic effects were graded according to the Common Terminology Criteria for Adverse Events version 3.0, and time to severe adverse event (grade 3 or higher) was retrospectively reviewed for every patient.
Continuous variables were compared using the Mann-Whitney U test, and categorical variables were compared using the chi-square test or Fisher's exact test. Overall survival was measured from day 1 of the first PIAF cycle to the date of death or last follow-up. Survival curves were generated using the Kaplan-Meier method, and differences between curves were evaluated using the log-rank test.
To identify prognostic factors for survival, a multivariate regression analysis was performed using the Cox proportional hazards model with backward elimination for variables with P < .1 in univariate analysis. To identify factors associated with an objective response to chemotherapy, a multivariate analysis was performed using the logistic regression model for clinical variables with P < .1 in univariate analysis. Analyses were performed with IBM SPSS software (version 19.0, SPSS Inc., Chicago, Ill). All statistical tests were 2-sided, and significance was set at P < .05.
Characteristics of the study population are summarized in Table 1. Compared with the conventional PIAF group, the modified PIAF group had a younger median age and a higher proportion of females. Positive hepatitis serology and liver cirrhosis were significantly less common in the modified PIAF group, and there was a trend toward better performance status in the modified PIAF group, reflecting the modified patient selection criteria and only limited to patients with well-controlled hepatitis, good liver function, and acceptable physical condition who underwent the modified PIAF treatment when the relative contraindication factors were present. There were no differences between the conventional PIAF group and the modified PIAF group in maximum tumor size, nodal involvement, presence of extrahepatic disease, or presence of major vascular invasion. However, number of tumors and serum alpha-fetoprotein levels were higher in the modified PIAF group.
Adverse Events, Objective Response Rate, and Rate of Conversion to Resectability
Outcomes of PIAF treatment are summarized in Table 2 in comparison with reported outcomes in previous studies. The median number of PIAF cycles was significantly higher in the modified PIAF group (4 versus 2; P = .049). Although the cumulative incidence of grade 3 or higher adverse events during PIAF treatment was similar in the 2 groups (58% for conventional PIAF versus 61% for modified PIAF), median number of cycles of PIAF before occurrence of a grade 3 or higher adverse event was significantly longer in the modified PIAF group (2 cycles versus 1 cycle; P = .003). As a result, the modified PIAF regimen was associated with a higher objective response rate (36% versus 15%; P = .013) and rate of conversion to curative surgery (33% versus 10%; P = .004).
Table 2. Treatment Details and Outcomes in Patients Treated with PIAF in the Current Study and Previous Studies
Figure 1 shows cumulative objective response rate during PIAF treatment. Median time to response was 210 days and the cumulative response rates were estimated as 4.2% at 8 weeks (2 cycles of PIAF), 15.4% at 12 weeks (3 cycles of PIAF), 22.3% at 16 weeks (4 cycles of PIAF), 31.2% at 20 weeks (5 cycles of PIAF), and 36.9% at 24 weeks (6 cycles of PIAF).
When the cutoff value for number of PIAF cycles was adjusted from 1 through more than 6, 5 or more PIAF cycles was associated with the highest odds ratio (OR) for objective response to chemotherapy (OR = 13.1; 95% confidence interval [CI] = 4.67-36.7; P < .01). Multivariate analysis confirmed that 5 or more PIAF cycles was the only predictor of objective response to PIAF treatment (OR = 15.6; 95% CI = 4.9-49.5; P < .001) (Table 3). The patients with conversion to curative surgery had a significantly higher median number of PIAF cycles (5 versus 2; P = .003) and better response compared with the patients whose disease remained unresectable, although the median pretreatment tumor size was similar (108 mm versus 100 mm in median, P = .78).
Table 3. Multivariate Analysis of Predictors of Objective Response
Among the 98 unresectable patients after PIAF treatment, 27 (36%) patients in the conventional PIAF group and 11 (50%) patients in the modified PIAF group received additional therapies. However, because of the high rate of severe adverse event and decreased performance status associated with PIAF treatment, available therapy was limited after completion of or failure from the treatment. Eventually, additional systemic therapy was performed in 23 (30%) patients in the conventional PIAF group and 7 (32%) patients in the modified PIAF group (P = .65). A regimen based on 5-FU or capecitabine was the most common regimens in either of the groups (87% versus 86%, P = 1). Sorafenib was used in only 1 patient after the modified PIAF treatment. TACE was selected in 5 (7%) patients in the conventional PIAF group and 4 (18%) patients in the modified PIAF group, respectively (P = .26). Radioembolization using yttrium-90 was performed for only 1 patient in the modified PIAF group.
At the time of this analysis, of the 84 patients in the conventional PIAF group, 78 had died, 5 had been lost to follow-up, and 1 was still alive, with a median follow-up period of 10.3 months (range, 1-211 months). Of the 33 patients in the modified PIAF group, 15 had died, 6 had been lost to follow-up, and 12 were still alive, with a median follow-up period of 15.3 months (range, 2-98 months). Intent-to-treat survival analysis revealed that median overall survival was significantly longer in the modified PIAF group (21.3 months versus 10.6 months; P = .002); the 3-year overall survival rate was 33.5% in the modified PIAF group versus 10.1% in the conventional PIAF group (Fig. 2A). When the groups were stratified according to whether patients underwent curative surgery, patients who underwent curative surgery had better survival regardless of which PIAF protocol was used (P < .0001) (Fig. 2B). Notably, among the patients who did not undergo curative surgery, patients treated with the modified PIAF regimen had better survival than those treated with the conventional PIAF regimen (19.0 months versus 10.1 months; P = .04). Median time to progression (TTP) for the 33 patients was 629 days (95% CI 124 days - not estimated), and median progression-free survival (PFS) was 64.4% at 6 months, and 55.2% at 1 year, while median TTP for the 84 patients was 201 days (95% CI 124-333 days), and PFS was 50.4% at 6 months, and 28.2% at 1year.
Multivariate analysis confirmed that ECOG performance status ≥ 2 (hazard ratio [HR] = 1.75; 95% CI = 1.04-2.93; P = .034) and positive hepatitis B serology (HR = 1.68; 95% CI = 1.08-2.59; P = .020) were associated with worse overall survival, whereas 5 or more cycles of PIAF therapy (HR = 0.56; 95% CI = 0.33-0.93; P = .024), curative surgical resection (HR = 0.15; 95% CI = 0.07-0.35; P < .001), and additional chemotherapy or transarterial chemoembolization after completion or failure of PIAF regimen (HR = 0.39; 95% CI = 0.25-0.61; P < .001) were associated with better overall survival (Table 4).
The median increase in serum alanine aminotransferase (ALT) level after initiation of PIAF treatment was significantly higher in patients with positive than in patients with negative hepatitis B serology (30 IU/L versus 1 IU/L; P = .008), and patients with positive hepatitis B serology had a higher incidence of hepatic dysfunction (ALT > 100 IU/L) during PIAF treatment (16 of 33 patients, 48%, versus 11 of 69 patients, 16%; P = .0006) according to the data for the 102 patients in whom both pre- and post-PIAF serum ALT levels were available. Furthermore, patients with ECOG performance status score of 2 or greater had a higher incidence of termination of PIAF treatment because of poor tolerance than did patients with ECOG performance score of 0 or 1 (6 of 21 patients, 29%, versus 15 of 81 patients, 16%; P = .23). When comparing objective response rate based on the PIAF protocols and presence of the relative contraindication factors (ie, positive hepatitis B serology, positive hepatitis C serology, ECOG performance status of 2 or greater, or liver cirrhosis), modified PIAF for patients without contraindication factors had the highest objective response rate (40%), whereas conventional PIAF for patients with the contraindication factors showed the lowest objective response rate (13%) (Fig. 3).
In this study, in patients with initially unresectable large HCC, the modified PIAF regimen in patients with no cirrhosis or hepatitis B was associated with improved tolerance, rate of response, resectability, and survival (Fig. 4). Receipt of 5 or more cycles of PIAF was associated with higher objective response rate and a rate of conversion to curative surgery of 33%. Five or more cycles of PIAF treatment was independently associated with improved survival in multivariate analysis (Table 4) regardless of whether surgical resection was performed. Taken together, these observations argue against the notion that HCC is uniformly chemoresistant and lend support for therapeutic approaches that take into account both HCC and any underlying liver disease. Our approach, if validated, could lead to establishing a successful neoadjuvant approach in potentially resectable large HCC tumors.
Table 4. Multivariate Analysis of Predictors of Survival
The emergence of chemotherapy in the 1950s has led to the availability of neoadjuvant systemic therapies for patients with advanced solid tumors. However, systemic cytotoxic therapies have demonstrated a very limited impact on the resectability of advanced HCC despite several studies showing reasonable response rates.[15-18] This is mainly because advanced underlying liver disease is associated with poor survival and tolerance of chemotherapy and is a contraindication to surgery in the first place. Recently, molecular characterization of hepatocarcinogenesis has led to recognition of defined aberrant signaling pathways, which has facilitated subsequent development of targeted agents as potential treatments for HCC, either alone or in combination with other therapies. In 2007, sorafenib was approved for treatment of patients with unresectable HCC based on possibly improving overall survival. However, sorafenib showed a very low response rate in randomized phase 3 trials[19, 20] and failed to downsize tumors to a resectable stage.
In a previous study, Leung et al reported that treatment with the PIAF regimen permitted curative surgery in 18% of patients who had initially unresectable HCC (Table 2). Although a subsequent phase 3 study by Yeo et al showed a surgical conversion rate of only 8% (Table 2), the study confirmed an overall objective rate of 21%, and excellent long-term survival was also confirmed in patients undergoing curative surgery after PIAF chemotherapy. In the current study, the patients treated with conventional PIAF had an objective response rate of 15% and a rate of conversion to curative surgery of 10%. However, selected patients with no cirrhosis who were treated with modified PIAF, similar to that of the phase 2 and 3 studies, had substantially higher rates of objective response (36%), and conversion to curative surgery (33%), and also had prolonged survival, even in patients with ultimately unresectable disease. Furthermore, the median size of tumor in patients achieving curative surgery was 10.8 cm, and this was substantially larger than the reported median size of HCC (4.7-7.0 cm) in randomized trials of TACE.[22, 23] These results suggest that neoadjuvant PIAF systemic therapy can offer a chance of cure even in selected patients with large HCC who are not good candidates for TACE.
Another noteworthy result of this study is that 5 or more cycles of PIAF appeared to be required for meaningful objective response and optimal survival, regardless of the PIAF protocol. This result suggests that the PIAF regimen must be tolerable in order to produce improved outcomes. The multivariate analysis showed that patients with chronic liver disease and worse performance status had worse prognosis. Indeed, these patients had a higher incidence of hepatotoxicity or termination of PIAF treatment because of toxic effect. With a carefully selected patient population and dose modification, tolerance of PIAF treatment was clearly improved, which resulted in higher number of cycles of treatment than in our conventional PIAF cohort or that of previous studies. These findings indicate that PIAF treatment should be limited to patients with no chronic liver disease and good performance status.
The limitations of this study include its retrospective nature and the difficulty of simple comparisons between the 2 cohorts because they differed with respect to both the PIAF protocol and patient selection. However, the modified PIAF dosing was based on results from other PIAF studies and the higher adverse events rate with the conventional PIAF dosing at our center. Furthermore, the current results are based on prospectively collected data, and the modified PIAF group showed substantially better outcomes compared with that of the conventional PIAF groups (Fig. 3) and previous PIAF studies that used very similar chemotherapy protocols with the modified PIAF group (Table 2).[8, 9] In addition, in the current era of effective anti–hepatitis B virus treatment, negative influence of hepatitis B on PIAF treatment could be minimized with adequate antiviral therapy. Therefore, further investigation about the effectiveness of PIAF therapy for patients with anti–hepatitis B virus treatment is needed. Another limitation is the potential small percentage of patients with no cirrhosis who would be good candidates for the PIAF regimen. However, our group and others observed a notable decreasing trend of the incidence of cirrhosis (55%-60%) and hepatitis (30%-47%) in HCC in recent years,[19, 24, 25] compared with the incidence in historic patients whose data were used to establish the HCC staging systems[26-30] (ie, hepatitis = 82.3%-100% and cirrhosis = 77%-100%). This is consistent with and complementary to multiple recent reports indicating the rising incidence of HCC cases that are not associated with hepatitis or cirrhosis, specifically among patients with nonalcoholic steatohepatitis and metabolic syndrome chronic liver diseases which can progress to HCC without apparent cirrhosis.[31-39]
In conclusion, in the current era of early goal-directed therapy and a focus on cost-effectiveness, it has become ever more important that we discriminate between patients with HCC who are at risk for clinically relevant adverse outcomes and mortality following specific therapy, such as treatment with the PIAF regimen, and patients who are at low risk for adverse outcomes, who stand to benefit most from therapy. Our proposed personalized therapy approach of neoadjuvant PIAF regimen to patients with no advanced chronic liver disease preserves the surgical (curative) option to patients who have a disease considered unresectable at diagnosis.
This research was supported in part by the National Institutes of Health through The University of Texas MD Anderson Cancer Center support grants CA170035-01 (to Dr. Kaseb), and CA106458-01 (to Dr. Hassan), and Cancer Center support grant CA016672, and by a grant from Schering-Plough to Dr. Patt.