To verify whether the PTPRD phosphatase-mut predictive efficiency was independent on other genes mutations, which were reported fundamental in response to ICIs, including TP53, EGFR, KRAS, STK11, and KEAP1, we tested whether PTPRD was co-mutated or exclusively mutated with such genes. In the NGS cohorts (Hellmann, Miao, N.Rizvi, and NCC cohort), there was no neither co-mutation nor exclusive mutation between PTPRD and these genes (all
p≥0.05, Additional file
2: Fig. S8a). In MSKCC-240 and MSKCC-350 cohort, which were sequenced by IMPACT panel, PTPRD was co-mutated with TP53 (
p<0.05), but not with other genes (all
p≥0.05, Additional file
2: Fig. S8b). In OAK/POPLAR cohort, which was sequenced by Foundation One panel, PTPRD was co-mutated with TP53 (
p<0.05), but not with other genes (all
p≥0.05, Additional file
2: Fig. S8c). In TCGA-LUAD cohort, which was sequenced by WES, there was no neither co-mutation nor exclusive mutation between PTPRD and these genes (Additional file
2: Fig. S8d).
As TP53 was found to be co-mutated with PTPRD, and EGFR, KRAS were two common mutations in lung adenocarcinoma and found to have great influences on responses to ICIs, we calculated PTPRD phosphatase-mut predictive efficiency in the TP53/EGFR/KRAS mutation/WT subgroups in different cohorts. In discovery cohort, no PTPRD phosphatase-mut was identified in the TP53 WT or EGFR mutation subgroup. In the WT KRAS subgroup (HR=0.21,
p=0.0078, Additional file
2: Fig. S8e), but not the mutant KRAS subgroup (HR=0.72,
p=0.75, Additional file
2: Fig. S8e), phosphatase-mut was a biomarker for PFS, but the
p value for interaction was 0.25 (Additional file
2: Fig. S8e). Regarding validation cohort 1, in the EGFR-WT and KRAS-WT subgroups, but not in the other subgroups, PTPRD phosphatase-mut was a biomarker for PFS (HR=0.52 (
p=0.04) and HR=0.47 (
p=0.01), respectively, Additional file
2: Fig. S8f). While the
p values for interaction of each gene were 0.5, 0.91, and 0.15 (Additional file
2: Fig. S8f). Regarding validation cohort 2, in the TP53-mut, EGFR-WT, and KRAS-WT subgroups, but not in the other subgroups, PTPRD phosphatase-mut was a biomarker for OS (HR=0.42 (
p=0.0034), HR=0.48 (
p=0.01), and HR=0.40 (
p=0.0063), respectively, Additional file
2: Fig. S8g). While the
p values for interaction of each gene were 0.79, 0.99, and 0.28, respectively (Additional file
2: Fig. S8g). In validation cohort 3, no PTPRD phosphatase-mut was identified in the EGFR-mut and KRAS-mut subgroups. PTPRD phosphatase-mut was a biomarker for PFS (HR=0.31 (
p=0.04), HR=0.30 (
p=0.03), HR=0.31 (
p=0.0049), HR=0.30 (
p=0.0046), respectively, Additional file
2: Fig. S8h) but not OS (Additional file
2: Fig. S8i) in both the TP53-WT and mutation subgroups. The
p value for the interaction of PTPRD phosphatase-mut and TP53 mutation/WT was 0.914 for PFS-HR (Additional file
2: Fig. S8h) and 0.515 for OS-HR (Additional file
2: Fig. S8i).
Taken together, although PTPRD was co-mutated with TP53 in MSKCC-240, MSKCC-359 and OAK/POPLAR cohort, PTPRD phosphatase-mut predictive efficiency was independent of TP53/EGFR/KRAS mutation.