Key New Studies in Gastric and Esophageal Cancer

In this issue of the Gastrointestinal (GI) Oncology Scan, we would first like to announce that Jeffrey Meyer will be rotating off of the GI editorial team. As one of the original associate editors to join during Editor-in-Chief Anthony Zietman’s reorganization in 2011, he has played a vital role in helping to shape and refine our review process, and we truly appreciate his commitment and dedication throughout the review process in ensuring that the highest-quality papers are accepted by this journal (1, 2). He will be focusing on new challenges and opportunities, and we bid him farewell and wish him the best. At the same time, we would like to welcome the newest member of the editorial team, Salma Jabbour, Associate Professor of Radiation Oncology at Rutgers Robert Wood Johnson Medical School and Rutgers Cancer Institute of New Jersey, who is a well-published and recognized leader in our field of GI radiation oncology. We look forward to the expertise that she brings to our group. Welcome Salma!

In step with our previous Oncology Scans covering recent findings in the journal literature with regard to GI cancers (3–7), our group has selected 5 important papers to highlight. The first 2 are randomized trials from Asia that help define the role of adjuvant therapy in patients with operable gastric cancer in the setting of a D2 lymph node dissection. The third study describes the profiles of different molecular subtypes of gastric cancer, which highlights how differences in recurrence pattern and subtype may affect the role of adjuvant therapy. We then switch gears to the PRODIGE 5 trial examining concurrent chemoradiation using standard cisplatin and 5-FU versus FOLFOX chemotherapy in the definitive setting for esophageal cancer. Finally, we selected an interesting study proposing an alternative grading system of liver function in patients with cirrhosis that may be very useful for how radiation oncologists select patients with primary liver cancer for radiation therapy, particularly with the increasing use of stereotactic body radiation therapy.

Park et al. Phase 3 trial to compare adjuvant chemotherapy with capecitabine and cisplatin versus concurrent chemoradiotherapy in gastric cancer: Final report of the adjuvant chemoradiotherapy in stomach tumors trial, including survival and subset analyses. J Clin Oncol 2015 Jan 5 [epub]. (8)

In this update of a phase 3 randomized trial from South Korea, 458 patients with stage IB to IV (M0) gastric cancer were randomized to 1 of 2 arms: (1) adjuvant cisplatin and capecitabine or (2) adjuvant cisplatin, capecitabine, and radiation therapy. All patients were randomized after undergoing curative surgical R0 resection with a D2 lymph node dissection. In the chemotherapy arm, patients received 6 cycles of capecitabine 1000 mg/m² twice daily for 2 weeks and cisplatin 60 mg/m², with each cycle lasting 3 weeks. In the chemoradiation arm, patients received a sandwich schedule of 2 cycles of capecitabine and cisplatin, then 45 Gy at 1.8 Gy/fraction with concurrent capecitabine 825 mg/m² twice daily, followed by 2 more cycles of capecitabine and cisplatin. Patients were stratified by pathologic stage (IB or II vs III or IV) and type of surgery (partial or total gastrectomy). The primary endpoint of the study was disease-free survival (DFS), and the accrual was set for 80% power to detect a hazard ratio (HR) of 1.45 (chemotherapy to chemoradiation) with a 2-sided α of 0.05.

After a median follow-up time of 7 years, the updated results revealed that there remained a nonsignificant improvement in DFS for chemoradiation versus chemotherapy, with an HR of 0.74 (P=.09) for DFS. The 5-year overall survival (OS) was 73% for chemotherapy and 75% for chemoradiation (P=.48). Local-regional recurrence was higher in the chemotherapy arm than in the chemoradiation arm (13% vs 7%, P=.0033). No difference was seen in distant metastases (27% vs 24%, P=.56). On subgroup analysis, patients with node-positive disease had significantly better 3-year DFS in the chemoradiation arm than did those in the chemotherapy arm (76% vs 72%, P=.04). Additionally, patients with intestinal-type cancer had better 3-year DFS with chemoradiation than with chemotherapy (94% vs 83%, P=.01). Of note, none of the biomarkers that were tested—HER2 (P=.99), EGFR (P=.59), MET (P=.099), or FGFR2 (P=.87)—were associated with DFS.

Noh et al. Adjuvant capecitabine plus oxaliplatin for gastric cancer after D2 gastrectomy (CLASSIC): 5-year follow-up of an open-label, randomised phase 3 trial. Lancet Oncol 2014;15:1389–1396. (9)

In this trial, 1035 patients with stage II to IIIB gastric cancer treated with gastrectomy and D2 lymph node dissection were randomly assigned to (1) observation or (2) adjuvant oxaliplatin and capecitabine for 6 months (eight 3-week cycles). Oxaliplatin was given at 130 mg/m² on day 1, and capecitabine was given at 1000 mg/m² twice daily for 2 weeks. Patients were required to have a minimum of 15 lymph nodes sampled. Accrual took place in South Korea, China, and Taiwan. The primary endpoint was 3-year DFS. In total, 515 patients were randomized to observation, and 520 patients were randomized to chemotherapy.

The median follow-up times were 62.6 months in the observation arm and 62.4 months in the chemotherapy arm. A total of 203 patients (39%) in the observation arm and 139 patients (27%) in the chemotherapy arm had disease recurrence, had a new gastric cancer, or died. Both DFS (HR 0.58, P<.0001) and OS (HR 0.66, P=.0015) were improved in the chemotherapy arm compared with the observation arm, and both of these improvements were observed when patients were analyzed on a per-protocol basis. The 5-year OS was 78% in the chemotherapy arm and 69% in the observation arm. DFS was improved with adjuvant chemotherapy in all disease stages (II, IIIA, IIIB), and OS was improved only among patients with stage II disease. Finally, on subgroup analysis, there was a significant interaction with T stage and adjuvant chemotherapy for OS. Patterns of failure analysis showed that 117 (23%) patients in the chemotherapy arm and 186 (36%) patients in the observation arm had recurrence. Of arm note, although distant metastases were the most common sites of failure, 27 (5%) patients in the chemotherapy arm and 52 (10%) patients in the observation arm had local-regional recurrence.

Comment:

Intergroup 0116 established adjuvant chemoradiation as the standard of care for gastric cancer after curative resection in the United States (10). There was a significant benefit in DFS and OS in patients who had adjuvant chemoradiation, which, based on patterns of failure data, could be mainly attributed to a reduction in local-regional recurrence. However, this trial has been criticized because most patients did not have what many consider to be an adequate lymph node dissection. Data from Asian centers strongly advocate for more extensive D2 lymph node dissections, even though a D2 dissection did not improve survival in Western patients in 2 randomized trials (11, 12). However, those results may be confounded by the increased morbidity seen with D2 dissection, and having an adequate number of lymph nodes removed remains an important surgical principle with gastric cancer. In the Intergroup trial, 54% of patients had a D0 dissection, and only 10% of patients had a D2 dissection. As a result, the patients in this trial may have been at higher risk for local-regional failure, thus inflating the benefit of chemoradiation. Critics have questioned the applicability of these data in the setting of Asian countries because of differences in both disease biology and extent of lymph node dissections.

The standard of care in many Asian centers has been adjuvant chemotherapy, which has been reinforced by the CLASSIC trial (9). In this study, patients randomized to chemotherapy had significantly better DFS and OS than those treated with surgery alone. The 5-year OS was quite good in both arms and superior to that in the Intergroup trial. It is likely that tumor biology between Asian and Western patients and earlier detection in Asian countries play important roles in this difference, as has been observed in other studies (13). In addition, all patients in this trial had a D2 dissection, which may also have led to the improved disease outcome. It is noteworthy that in both arms, the rate of local-regional recurrence was quite low at 5% to 10%, implying a limited role for adjuvant radiation after extensive lymph node dissection. The results from the CLASSIC trial are consistent with the ACTS-GS (14), which showed that adjuvant S-1 chemotherapy improved survival compared with surgery alone in Japanese patients with stage II and III gastric cancer after D2 dissection. Trials in Western countries have shown conflicting results regarding the benefit of adjuvant chemotherapy (15, 16), although a meta-analysis confirmed a small but significant improvement in OS and progression-free survival (PFS) (17). Altogether, the role of adjuvant chemotherapy seems clear, but the optimal adjuvant regimen and duration of therapy are still unknown.

The ARTIST trial examines the role of radiation in the setting of a D2 lymph node dissection and adjuvant chemotherapy in Asian patients (8). This important trial addresses many of the critical concerns raised about the Intergroup trial. In this randomized phase 3 trial, all patients received either chemoradiation sandwiched between doublet chemotherapy for 4 total cycles versus doublet chemotherapy alone for 6 total cycles. The updated results of this trial found a nonsignificant benefit in DFS (P=.09) with the addition of chemoradiation but no difference in OS (P=.52). However, on unplanned subgroup analysis, a significant improvement in DFS with chemoradiation was observed in patients with node-positive disease and in patients with intestinal-subtype histology. These results complement 2 other trials from Asia that showed an improvement with adjuvant chemoradiation versus chemotherapy alone after D2 dissection (18, 19). One trial demonstrated an improvement in relapse-free survival but not OS in 380 patients (19). The other trial only randomized 90 patients with stage III and IV(M0) disease but showed improvement in local-regional relapse-free survival and DFS in the stage III patients, patients with primarily node-positive disease but with more favorable prognosis than those with stage IV(M0) disease (18).

The results of these trials highlight the important role that radiation may play in specific subgroups of patients, particularly those with a higher burden of local-regional disease. In addition, histology seems also to play an important role in determining which patients may benefit from which therapy. Diffuse-type gastric cancer has a higher propensity for metastatic spread (20); therefore, these patients may not benefit as much from additional local-regional therapy. Similar to the ARTIST trial, the Intergroup 0116 trial showed that patients with diffuse gastric cancer did not benefit from adjuvant radiation.

The authors speculate that the overall results of the ARTIST trial may have failed to show a benefit for chemoradiation because of the large proportion of patients with earlier-stage disease (60% stage IB and II) who may have more favorable outcomes. In addition, they highlight that approximately 60% of patients had diffuse histology in this group, which also likely had an impact on the overall results, given the apparent lack of benefit of radiation for this group.

Taken together, these 2 trials, the CLASSIC trial and the ARTIST trial, help to further refine the role of adjuvant therapy for gastric cancer beyond the Intergroup 0116 treatment. Chemotherapy appears to be indicated for all patients with stage IB to stage IV disease, particularly those with diffuse-type histology, although the optimal regimen and duration of chemotherapy have yet to be determined. Adjuvant chemoradiotherapy after D2 lymph node dissection seems most appropriate for patients with lymph node—positive disease or intestinal-type histology and is also supported by other series (21). Further investigation is needed to test this treatment approach, and the ARTIST 2 trial is currently under way testing adjuvant chemotherapy for 1 year versus chemotherapy for 6 months versus chemoradiation in patients with node-positive gastric cancer. The other important caveat that needs to be answered is the applicability of these results to patients in Western countries.

Cristescu et al. Molecular analysis of gastric cancer identifies subtypes associated with distinct clinical outcomes. Nature Med 2015;21:449–456. (22)

In this report, the Asian Cancer Research Group (ACRG) analyzed gene expression profiles from 300 gastrectomy specimens from Samsung Medical Center to derive 4 distinct gastric cancer molecular subtypes found to be associated with survival and recurrence patterns after surgery. Tumors classified as having microsatellite instability (MSI) were hypermutated, intestinal-subtype tumors (60%), were most often located in the antrum (75%), were more commonly diagnosed as stage I or II (>50%), and had the best overall prognosis. Mesenchymal-like type tumors (MSS/EMT) were mostly diffuse-type tumors (80%), showed signet ring cells (44%), occurred at a younger age, were more often diagnosed as stage III or IV, and had the worst prognosis. Two intermediate-risk subtypes were further stratified by TP53 status, defined as MSS/TP53− and MSS/TP53+, with the former having a relatively worse prognosis. The presence of these subtypes and their associations with survival were validated in independent datasets including a second Samsung Medical Center cohort, a Singaporean cohort, and the Cancer Genome Atlas (TCGA) gastric cohort, which totaled >650 samples. The risk of recurrence was highest for MSS/EMT (63%), and recurrence was most likely with peritoneal seeding (64%) compared with other subtypes. Meanwhile, liver-limited metastases were more common in the MSI (23%) and MSS/TP53− (21%) subtypes compared with MSS/EMT (4.6%) and MSS/TP53+ (8%). The proposed classification system by the ACRG was compared with the 4 TCGA subtypes of EBV+, MSI, genome stable, and chromosomal instability and was found to have slight differences in the cohort, molecular mechanism, driver gene, and prognosis association. Key gene amplification and somatic mutations were provided for each subtype.

Comment:

This study is a nice complement to the ARTIST trial, which showed the impact that histology has on survival and response to adjuvant therapy. Molecular profiling and DNA sequencing of tumors represents the future of “personalized” cancer care, and there has been much interest and progress made in many cancers, including gastric cancer. For example, trastuzumab is already in clinical use for patients with HER2 overexpressing gastric cancer, having shown improvement in survival in the metastatic setting (23). By identifying distinct molecular subtypes with different prognoses and patterns of recurrence, this study by Cristescu et al (22) takes an important step towards furthering our understanding of the link between cancer genotype and clinical phenotype. These provocative results have important implications on tailoring therapy based on the molecular profile. As shown in the ARTIST trial, patients with diffuse histology did not benefit from adjuvant radiation therapy, and histologic type is likely to play an expanding role in determining the optimal adjuvant therapy. Further customizing of treatment based on molecular profiling represents the next logical step in this evolving field, and an improved understanding of how the genotype determines clinical outcome is essential, especially with the rapid advancements made in molecularly-targeted therapies. In their discussion, the authors provide an excellent overview of the clinical implications of their work and suggestions for future research directions.

Conroy et al. A Fédération Francophone de Cancérologie Digestive and UNICANCER-GI Group. Definitive chemoradiotherapy with FOLFOX versus fluorouracil and cisplatin in patients with oesophageal cancer (PRODIGE5/ACCORD17): Final results of a randomised, phase 2/3 trial. Lancet Oncol 2014;15:305–314. (24)

In this randomized phase 2/3 trial, 267 patients with stage I to IVA (American Joint Committee on Cancer 6th edition) esophageal cancer (squamous cell carcinoma, adenocarcinoma, or adenosquamous carcinoma) were enrolled to either FOLFOX or cisplatin with 5-FU chemotherapy given concurrently with radiation therapy and adjuvantly. Radiation therapy was prescribed to 50 Gy at 2 Gy per fraction. The clinical target volume (CTV) included a 3- to 5-cm proximal and distal margin and a 2-cm radial margin. For tumors of the upper third of the esophagus, the supraclavicular nodes were treated to 40 Gy. The celiac nodes were included for tumors of the lower third of the esophagus. The FOLFOX arm initially consisted of 6 cycles given every 2 weeks, oxaliplatin 85 mg/m² on day 1, leucovorin 200 mg/m² on day 1, and 5-FU 400 mg/m² bolus on days 1 and 2 followed by continuous-infusion 5-FU at 600 mg/m² on days 1 and 2. The first 3 cycles were given concurrently with radiation therapy. Because of the concern about neutropenia, the protocol amended the dose of 5-FU. The cisplatin and 5-FU arm consisted of 4 total cycles of cisplatin 75 mg/m² on day 1 and 5-FU 1000 mg/m² continuous infusion on days 1 to 4. The first 2 cycles were given concurrently with radiation at 4-week intervals, and the 2 adjuvant cycles were given 3 weeks apart. The trial was designed to show a 20% increase in PFS at 3 years with FOLFOX over cisplatin/5-FU (from 30% to 50%) with 90% power and an α of 0.05.

In total, 134 patients were assigned to FOLFOX and 133 to cisplatin/5-FU. All but 2 patients in the FOLFOX group and 1 in the cisplatin/5-FU group received ≥90% of the prescribed radiation dose. Chemotherapy dose reduction due to toxicity was needed in 34% and 27% of patients in the FOLFOX and cisplatin/5-FU arms, respectively (P=.19). Concurrent chemoradiation was delivered as planned in 93% of patients in the FOLFOX group and 97% in the cisplatin/5-FU group, and 71% of patients received all 6 cycles of FOLFOX compared with 76% of patients who received all 4 cycles of cisplatin/5-FU. With a median follow-up time of 25.3 months for all enrolled patients, the median PFS was 9.7 months in the FOLFOX arm and 9.4 months in the cisplatin/5-FU arm (HR 0.93, 95% CI 0.7–1.24, P=.64). The 3-year PFS in the FOLFOX arm was 18.2% compared with 17.4% in the cisplatin/5-FU arm. No difference in OS was seen between the 2 arms, 20.2 months in the FOLFOX arm versus 17.5 months in the cisplatin/5-FU arm (HR 0.94, 95% CI 0.68–1.29, P=.70), and the 3-year OS was 19.9% versus 26.9%, respectively. No difference was seen in complete (44% vs 43%), partial (22% vs 22%), and overall response rates (67% vs 65%) between the FOLFOX and cisplatin/5-FU arms. Similarly, no difference was seen in primary tumor progression (26% vs 25%), nodal progression (13% vs 17%), or distant progression (32% vs 27%) between the FOLFOX and cisplatin/5-FU arms. Overall, no differences in grade 3 or 4 toxicity rates were noted between the 2 arms, although the FOLFOX arm had significantly more paresthesias (47% vs 2%, P<.0001), sensory neuropathy (18% vs 1%, P<.0001), and elevation of liver enzymes (11% vs 2%, P=.002) compared with the cisplatin arm, and the cisplatin arm had higher rates of elevated creatinine (12% vs 3%, P=.007), mucositis (32% vs 27%, P=.011), and alopecia (9% vs 2%, P=.005).

Comment:

This trial was selected here because of its importance in potentially altering the standard of care for patients undergoing chemoradiation therapy as definitive treatment for esophageal cancer. Toxicities with combined modality chemoradiation therapy for esophageal cancer can be significant, and cisplatin plus 5-FU has long been considered the standard of care regimen used concurrently with radiation. This trial highlights the differences in toxicity profiles yet similar outcomes between FOLFOX and cisplatin plus 5-FU with concurrent radiation therapy (24). On the basis of this study, oncologists may now have additional discretion to tailor the chemotherapy to the patient’s potential tolerance in the definitive setting. In comparing the toxicity profiles, FOLFOX was demonstrated to reduce renal complications, mucositis, and alopecia compared with cisplatin plus 5-FU. The full course of chemoradiation therapy was fully completed by 67% of patients assigned to FOLFOX and similarly completed by 72% of the cisplatin/5-FU group.

The chemotherapy regimen is an important determinant of the overall toxicity profile of chemoradiation therapy, and there are 2 ongoing phase 2 trials which have abandoned the cisplatin plus 5-FU combination with radiation therapy. The phase 2 NEOSCOPE trial in the United Kingdom is enrolling patients with resectable gastroesophageal junction carcinoma to evaluate the utility of 2 cycles of induction chemotherapy with oxaliplatin plus capecitabine versus carboplatin plus paclitaxel followed by chemoradiation and surgery (25). Carboplatin plus paclitaxel concurrent with radiation therapy is a relatively well-tolerated and effective regimen, as was recently seen in the CROSS trial (26). Likewise, the CALGB 80803, also for gastroesophageal junction adenocarcinomas planned for surgery, is a recently-completed phase 2 trial that incorporated PET scan responses to individualize the chemotherapy regimen of FOLFOX or carboplatin/paclitaxel after induction chemotherapy and was based on the MUNICON trials (27, 28).

Also, both the PRODIGE and CALGB trials have collected quality of life data, which will help oncologists to understand the impact of these therapies on the daily lives of patients. The CALGB 80803 trial and NEOSCOPE trial will provide additional comparative information about the toxicity and efficacy of FOLFOX and carboplatin/paclitaxel concurrently with radiation therapy. Whereas carboplatin plus paclitaxel has replaced cisplatin/5-FU as the standard of care in the preoperative setting since the CROSS trial results were reported, it is likely, based on recent and upcoming data, that either FOLFOX or carboplatin/paclitaxel may do the same in the definitive setting.

Johnson et al. Assessment of liver function in patients with hepatocellular carcinoma: A new evidence-based approach: The ALBI grade. J Clin Oncol 2015;33:550–558. (29)

In this study, patients with hepatocellular carcinoma (HCC) from several international high-volume HCC centers were pooled to determine a model for liver function using only serum albumin and bilirubin, which are objective laboratory data, to predict survival and compare that model with the existing Child-Turcotte-Pugh (CTP) classification. In total, 8568 patients were included from the United Kingdom (n=1356), Spain (n=834), Japan (2599), China (n=1112), and the United States (n=509) along with patients from clinical trials (n=1132), patients who underwent resection (n=525), and patients with cirrhosis without HCC (n=501). The Japanese subset was randomly divided into a training cohort (n=1313) and a validation cohort (n=1286). The predictive model was based on the Japanese training set and was subsequently tested against all groups. The model used 3 risk categories: low (lowest 25th percentile risk), medium (25th to 90th percentile risk), and high (highest 10th percentile risk).

On multivariate analysis, tumor size, vascular invasion, albumin, log₁₀ bilirubin, total number, age, and gender were independently associated with survival. When tumor-related factors were removed, serum albumin and log₁₀ bilirubin were also associated with survival, and these 2 parameters were chosen for inclusion into the predictive model called the ALBI grade because they relate to liver function. The ALBI grade is a 3-point scale based on the score calculated from the following formula:

$$\text{Linear predictor}=){\text{log}}_{10}\text{bilirubin}\times 0.66)+(\text{albumin}\times -0.085)$$

Bilirubin is in μmol/L and albumin is in g/L. An ALBI grade 1 is assigned to a score of ≤−2.60, an ALBI grade 2 is assigned to a score of −2.6 to ≤−1.39, and ALBI grade 3 is assigned to a score >−1.39. Application of the ALBI grade to all groups showed similar discrimination of survival as CTP score. For patients within the CTP A classification, the ALBI grade separated patients from the Japanese, European, United States, and Chinese datasets into 2 different prognostic categories. A similar finding was seen for patients with CTP A cirrhosis receiving sorafenib. In patients with cirrhosis alone without HCC, the ALBI grade was able to distinguish 3 separate prognostic categories. Finally, in patients who underwent resection, ALBI grades 1 and 2 separated patients into 2 distinct prognostic groups, whereas CTP A and B class did not. The authors also provide a nomogram to determine the ALBI grade.

Comment:

The prognosis of a patient with HCC is a function of the biological nature of the primary tumor and the degree of underlying liver dysfunction. At least 8 staging systems have been developed specifically for HCC, including the Barcelona Clinic Liver Cancer (BCLC), Okuda, and Cancer of Liver Italian Program (CLIP) systems (30). In contrast, liver function is commonly assessed using the CTP system, which originally was developed in the 1960s to 1970s to help estimate survival times for patients with cirrhosis who were undergoing surgery for variceal bleeding (31, 32). The article highlighted here by Johnson et al (29) identifies the ABLI system as a meaningful alternative to the CTP system assessment of liver function based on purely objective data and inherently provides insight into the patient’s HCC.

Historically, the CTP system has been extrapolated from its original use also to estimate the tolerance of the cirrhotic liver to various liver-directed and systemic treatments for HCC. It is based on 5 factors: albumin and bilirubin levels, international normalized ratio (INR) values, and the degree of ascites and encephalopathy. The primary limitation of the CTP calculation is the clinician’s subjectivity in estimating ascites and encephalopathy, the varying success of medical management, and the effect of the HCC on the development of ascites. Often, the INR values are not available, and in 1 international HCC registry study, approximately a quarter of patients were unable to be assigned a CTP score because of missing INR values (33). Nevertheless, the CTP score’s clarity of calculation, stratification, and overall accuracy has led to its widespread use for decades. The ALBI scoring system may be better applied to current practice to evaluate management options, including systemic therapy and local treatments, such as surgery. In the future, with additional validation, it may help to predict the optimal candidates for radiation therapy.

Stereotactic ablative body radiation (SABR/SBRT) is currently a therapeutic option for patients with nonoperative HCC. In the setting of SABR for HCC, the de facto standard of care is to treat CTP grade A patients because SABR can be accomplished safely and effectively in this group (34), forming the basis of the current NRG/RTOG 1112 phase 3 clinical trial comparing sorafenib alone with SABR followed by sorafenib. Evaluation of the CTP B grouping shows a potential for liver function decline after SABR (35); however, patients with small HCCs and modestly impaired CTP B7 liver function may, in fact, safely tolerate SABR (36, 37). Others have demonstrated a decline in CTP class after SABR, with V25 > 36% being associated with a 4-fold increase in the incidence of Child-Pugh class decline (38). Another study evaluating the threshold dose of the liver as seen by MRI proposed a dose of 30 Gy for CTP A and 25 Gy for CTP B groups in 5 fractions to predict potential loss of liver tissue after SABR (39).

Although not its original intent, the CTP scoring system is currently the most commonly used method to stratify patients into risk categories of liver tolerance to radiation therapy, including SABR. Although we do not expect that to change any time soon, given the subjectivity of some of the criteria, we are intrigued with the potential of the ALBI system, which appears to be quite well validated here in a very large population of Asian and non-Asian patients for predicting survival as well as, or possibly better than, the CTP system. Its appeal lies in its simplicity of using only 2 objective factors from the CTP system while eliminating the 2 subjective factors—degree of ascites and encephalopathy—which thus allows for more reproducible estimates of liver function. For now, the ABLI grade may serve as a simpler and possibly more discriminating predictor of long-term prognosis with surgery and sorafenib, but with additional investigation, it may also be used for other liver-directed therapies such as SABR. We therefore would encourage future studies using radiation therapy or SABR to incorporate the ALBI system to better establish its role as a means to predict liver tolerance and risk of toxicity to radiation.