In:
Journal of Clinical Oncology, American Society of Clinical Oncology (ASCO), Vol. 39, No. 3_suppl ( 2021-01-20), p. 87-87
Abstract:
87 Background: Anti-PD-(L)1-based therapy yielded unprecedented efficacy in patients with MSI-high mCRC. A relevant proportion of subjects may achieve long-term disease control when receiving ICIs, particularly anti-PD-1 plus anti-CTLA-4 combo. However, some patients still experience treatment refractoriness or short-term clinical benefit. Methods: We analyzed data of 163 patients with MSI-high mCRC treated with anti-PD-1 +/- anti-CTLA-4 agents. The endpoint was progression-free survival (PFS); multivariable analyses were performed using a cure model (Othus et al, CCR 2012), which allows to test which factors, including ICI type, could identify patients achieving long-term disease control. To account for biases consistent with non-random ICI assignment, we estimated a propensity score (covariates: ECOG PS, age, sex, primary tumor location, its resection, adjuvant treatment, synchronous presentation of mets, mucinous histotype, RAS/BRAF status, n. prior treatment lines, previous chemo, n. metastatic sites, presence of peritoneal, lung, liver, bone, brain, nodal mets), and then in the cure model we applied an inverse-probability-of-treatment-weight (IPTW) based on propensity score. A beforehand variable selection was operated using a random survival forest (RSF) model (RSF covariates: all the propensity score covariates plus ICI line and ICI type, platelets count (PLT), NLR, LDH), where we introduced a weight system in order to adjust variable selection net of ICI type. Results: RFS selected 5 variables: ICI type, ECOG PS, NLR, PLT and N. prior lines, the combination of which allowed to estimate the cure probability: the higher the probability the greater the chance of achieving long-term disease control. The Table shows the odds ratio estimates, defining the chance of being “cured” according to the values of the selected variables. ICI type was significantly associated with long-term disease control: patients treated with anti CTLA-4-combo achieved the highest chance of being cured, 3.41 times greater than the chance of patients treated with anti PD-1 mono. The median cure probability was as higher as 70.1% (IQR: 51.9-81.0%) with CTLA-4-combo vs 34.8% (IQR: 18.4-51.9%) with anti-PD-1 mono; the probability within treatment groups was varying according to ECOG PS, NLR, PLT and n. prior lines. Conclusions: Based on 5 variables including the ICI treatment type, a nomogram was built in order to estimate the cure probability in patients with MSI-high mCRC and potentially assist clinicians in their clinical practice. [Table: see text]
Type of Medium:
Online Resource
ISSN:
0732-183X
,
1527-7755
DOI:
10.1200/JCO.2021.39.3_suppl.87
Language:
English
Publisher:
American Society of Clinical Oncology (ASCO)
Publication Date:
2021
detail.hit.zdb_id:
2005181-5
Bookmarklink