Elevated tumor markers for monitoring tumor response to immunotherapy.

Background: As the immune-related response evaluation criteria in solid tumors (irRECIST) by imaging greatly underestimated the objective response to immunotherapy, we established the response evaluation criteria in solid tumors based on tumor markers (RecistTM) to explore whether RecistTM can compensate for the deficiencies of the irRECIST criteria.

Methods: This was an observational study, which consisted of two parts. The first part (Group A) was a retrospective study including the patients with malignant solid tumors. The second part (Group B) was a prospective study, which were EGFR-negative and ALK-negative patients with stage IIIB-IV non-small cell lung cancer receiving first-line treatment. From January 2017 to September 2020, one hundred and ten patients with a three-time increase in tumor markers receiving immunotherapy were recruited. The treatment response to immunotherapy was evaluated by irRECIST and RecistTM. Efficacy, overall survival (OS), first evaluation time and earliest response time under the different evaluation criteria were compared by statistics.

Findings: The treatment response evaluated by the RecistTM criteria was not consistent with that evaluated by the irRECIST criteria (Kappa = 0.386,  < 0.001). RecistTM had a higher completed response (CR) rate compared to irRECIST criteria (20.9% vs 1.8%,  < 0.001). The earliest response time under the RecistTM criteria was 3.42 weeks earlier than that under the irRECIST criteria ( = -5.233,  < 0.001). There were significant differences in median OS between tumor marker-related complete response (tmCR) and tumor marker-related partial response (tmPR), as well as between tmPR and tumor marker-related stable disease (tmSD) (χ = 15.572,  < 0.001; χ = 7.720,  = 0.005), but not between tmSD and tumor marker-related progressive disease (tmPD) (χ = 1.596,  = 0.206). When applying both criteria together, for patients with immune-related CR / immune-related PR (irCR/irPR) ( = 54) under irRECIST criteria, there was a significant difference in median OS between achieving tmCR ( = 22) and tmPR ( = 32) (χ = 14.011,  < 0.001). RecistTM criteria can predict 1-year and 2-year OS more accurately than irRECIST criteria (AUCs:0.862 vs 0.552, 0.649 vs 0.521, respectively;both  < 0.001). In RecistTM, 4 patients had been observed with pseudoprogression in tumor markers.

Interpretation: The RecistTM criteria could effectively distinguish CR, PR, and SD, which may help resolve the shortcomings of the RECIST criteria in evaluating the treatment response to immunotherapy, especially in assessing whether patients can achieve deep or even complete response as soon as possible.

Funding: This work was supported by the Key projects of Chongqing Health and Family Planning Commission (to Xueqin Yang, 2019ZDXM011).