Programmed cell death protein 1 (PD-1) expressed on T lymphocytes and activation of the receptors by their ligands (PD-L1) prevents the reaction of T cells to tumor cells and thereby antitumor immunity (1–3). In recent years, additional antibodies targeting the PD-1/PD-L1 have emerged as a promising treatment for inhibiting the progression, relapse, and metastasis of advanced hepatocellular carcinoma (HCC) (4). The initial effect of an PD-1/PD-L1 inhibitor can result in an increase in tumor diameter caused by immune cell infiltration (5). The coexistence of immune cell infiltration and chronic inflammation in HCC might develop an atypical response pattern and patients may initially meet the criteria for progressive disease (PD) but later show stable or reduced tumor burden (6). This can result in failure of the conventional criteria, named Response Evaluation Criteria in Solid Tumors 1.1 (RECIST 1.1) (7), to capture the atypical patterns of tumor response. To address this issue, modified RECIST (mRECIST) (8, 9) and immunotherapy modified Response Evaluation Criteria in Solid Tumors (iRECIST) (10) were introduced, aiming to assist with discrimination between PD and pseudoprogression (PSPD). mRECIST only measures the arterially enhanced parts of the HCC target lesions. iRECIST is based on RECIST 1.1. The major difference between RECIST 1.1 and iRECIST is that the lesion categorized as PD in RECIST 1.1 requires verification by subsequent examinations to distinguish PD from PSPD. A study assessed the efficacy of nivolumab in patients with advanced HCC who achieved an overall response rate (ORR) of 18% by mRECIST and 14% by RECIST 1.1 following second-line treatment (11). Simultaneously, many studies have shown that the conventional criteria (RECIST 1.1) underestimated the benefits of immunotherapy in solid tumors, which may cause an early discontinuation of immune treatment (12–14). In addition, the ORR by mRECIST was regarded as a dependent predictor of overall survival (OS) in HCC treated with targeted therapy, meaning that responders survive significantly longer than nonresponders (15–19). However, there was no consensus on which response evaluation criteria were superior in immunotherapy of HCC, which led to complexities in patient management in different clinical trials. Additionally, the comparisons between the current standard (RECIST 1.1, iRECIST, and mRECIST) assessments are inadequate in advanced HCC. The evidence regarding the association of OS and response assessment by RECIST 1.1, iRECIST, and mRECIST in HCC treated with PD-1 inhibitor is scarce, especially in real-world settings.

The purpose of this study was to compare the response assessments using RECIST 1.1, mRECIST, and iRECIST and to evaluate the correlation between survival outcome and response assessment achieved by these criteria in HCC treated with PD-1 inhibitor.

Recent immune therapies, including immune checkpoint inhibitors, have shown encouraging clinical results in patients with advanced HCC (20, 21). Programmed death receptor-1 (PD-1)/PD-1 ligand (PD-L1) inhibitors, such as nivolumab and pembrolizumab, have been conditionally approved by the US Food and Drug Administration for second-line treatment of advanced HCC (4, 22). In addition, a number of criteria, including traditional evaluation criteria (RECIST 1.1) and immune-related criteria (iRECIST and mRECIST), have been reported to assess the efficacy of immunotherapy. However, it is unclear which criteria are best for evaluating immune-related response. In addition, the correlations between survival outcome and response assessment achieved by conventional and immunotherapy-modified methods in patients with HCC are also unknown.

This study explored a longer ORR and a greater number of BOR by mRECIST compared with the other criteria. A significant difference between responders and nonresponders assessed by mRECIST (p < 0.05) was observed. Our study showed a significantly longer OS and PFS in responders than nonresponders assessed by mRECIST than traditional evaluation criteria (p < 0.05).

The tumor assessment by RECIST1.1 is more rigorous and accurate than mRECIST. Patients who have achieved complete tumor remission by mRECIST were responding to immunotherapy actually. Thus, immunotherapy for those patients will be expected to extend the patient’s OS. This means that the mRECIST standard actually prolongs the survival period of some patients who “cannot benefit from immunotherapy”. Furthermore, the intrareader variability of mRECIST was lower than that of the other criteria evaluated in this study and may explain why mRECIST is not widely used in clinical applications and indicates that mRECIST needs to be improved.

Previous studies have shown that conventional criteria underestimated the ORR for immunotherapy by up to 15% (23–26). Simultaneously, it achieved an ORR of 18% by mRECIST and 14% by RECIST 1.1 in second-line treatment of advanced HCC (11). Thus, the conventional criteria (RECIST 1.1) may underestimate the effectiveness of immunotherapy. Our study found that the change in ORR between mRECIST (12 of 50, 24%) and RECIST 1.1 (8 of 50, 16%) was 8% in patients with advanced HCC treated with PD-1 inhibitor. The relatively small sample size in the present research might have affected this result. A recent study has shown that a high ORR to nivolumab in phase II trials was associated with prolonged OS (11). Many studies have indicated that mRECIST criteria in HCC improved the sensitivity to quantify tumor response with targeted therapies (19) and patients who achieved a response on sorafenib had longer survival than nonresponders (27, 28). However, there is little research on the relationship between tumor response and survival in HCC patients treated with immunotherapy. Our study found that OS and PFS of the immunotherapy response group were significantly longer than those in the nonresponse group, which was classified by mRECIST (p < 0.05). Hence, our novel findings may suggest that mRECIST is better than RECIST 1.1 and iRECIST in determining immunotherapy response.

The mOS and mPFS of our study population were 20.9 ± 12.5 months and 11.5 ± 10.2 months based on follow-up data. The OS and PFS in our research were longer than previous studies for an immunotherapy monotherapy (5, 29). According to the Child-Pugh and ECOG scores of the patients included in our study, it showed that the patients were in good physical condition. The patients we included had fewer number that occurred macrovascular invasion (16% vs.32%) and less number of extrahepatic metastases (64% vs.71%) compared with CheckMate 040 study. In addition, most patients received further systemic treatments after anti-PD-1 discontinuation. These factors may lead to differences in patient survival analysis between our studies and other clinical studies.

This study attempted to demonstrate the intrareader variability of mRECIST in immunotherapy of HCC for the first time. The concordance rate according to Cohen k was rather low compared with other papers (30). We suggest that the reason for the fair intrareader variability may be caused by the mRECIST measurement method which included measurement of the longest viable tumor diameter for the assessment of response. It is challenging for doctors to accurately measure the longest diameter of the viable tumor in lesions showing partial internal necrosis. Soft tissue resolution of CT is lower than MRI, leading to comparatively low disease detection and measurement sensitivity. Comparatively low disease detection and measurement sensitivity of CT could lead to deviation when determining tumor boundary and measuring tumor diameter. Furthermore, there may be a slight difference in the timing of each CT scan. The drugs used by patients in the two studies were different (PD-1 vs. sorafenib). Special therapeutic responses to tumor immunoassay inhibitor treatment, such as pseudo-progression and hyperprogress, etc., made it more difficult for radiologists with different qualifications. Thus, there is an urgent need for a more reasonable mRECIST assessment that provides a reliable method for assessing tumor response in HCC clinical trials. Quantitative imaging parameters related to immune efficacy evaluation are expected to be added to the mRECIST. Hence, our novel findings may indicate a disadvantage of mRECIST and provide new insights into mRECIST.

There were some limitations in this study. On the one hand, it was a retrospective study from a single center, resulting in recruitment bias. On the other hand, due to the limited number of patients treated with PD-1/PD-L1 inhibitor, further large-scale prospective studies are required to validate the results and examine the clinical utility of mRECIST in advanced HCC patients.

In conclusion, the immunotherapy-modified assessment method (mRECIST) has a powerful ability to discriminate between responders and nonresponders and shows significantly longer OS and PFS in responders than in nonresponders. However, it requires further improvements in order for it to be widely used in the clinical evaluation of immunotherapy in HCC.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding authors.

The studies involving human participants were reviewed and approved by Harbin Medical University Cancer Hospital. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements. Written informed consent was not obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

Conceived and designed the analysis: TZ and YZ. Collected the data: MZ, YS, YX, JN, FW, and YH. Performed the analysis: MZ, YZ, CZ, SL, and YW. Drafting of the manuscript: TZ and MZ. Wrote the paper: MZ, TZ, CZ, and JN.

This study was supported by “Tou Yan Action” of Heilongjiang province, the National Natural Scientific Foundation of China (No. 81472322, No. 81872435, No. 81672930, and No. U20A20377); the Provincial Natural Science Foundation Outstanding Youth Project (JQ2019H003); the National Youth Talent Support Program for TZ; the Natural Science Foundation of Heilongjiang Province (LC201437/H1617); the Fok Ying Tung Education Foundation (No. 151037); Scientific Research Project of the Health Planning Committee of Heilongjiang (2017130); the Academician Yu Weihan Outstanding Youth Foundation of Harbin Medical University for TZ; and the HaiYan Funds of Harbin Medical University for YZ (JJZD2020-17).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.