High familial burden of cancer correlates with improved outcome from immunotherapy in patients with NSCLC independent of somatic DNA damage response gene status

Family history of cancer (FHC) is a hallmark of cancer risk and an independent predictor of outcome, albeit with uncertain biologic foundations. We previously showed that FHC-high patients experienced prolonged overall (OS) and progression-free survival (PFS) following PD-1/PD-L1 checkpoint inhibitors. To validate our findings in patients with NSCLC, we evaluated two multicenter cohorts of patients with metastatic NSCLC receiving either first-line pembrolizumab or chemotherapy. From each cohort, 607 patients were randomly case–control matched accounting for FHC, age, performance status, and disease burden. Compared to FHC-low/negative, FHC-high patients experienced longer OS (HR 0.67 [95% CI 0.46–0.95], p = 0.0281), PFS (HR 0.65 [95% CI 0.48–0.89]; p = 0.0074) and higher disease control rates (DCR, 86.4% vs 67.5%, p = 0.0096), within the pembrolizumab cohort. No significant associations were found between FHC and OS/PFS/DCR within the chemotherapy cohort. We explored the association between FHC and somatic DNA damage response (DDR) gene alterations as underlying mechanism to our findings in a parallel cohort of 118 NSCLC, 16.9% of whom were FHC-high. The prevalence of ≥ 1 somatic DDR gene mutation was 20% and 24.5% (p = 0.6684) in FHC-high vs. FHC-low/negative, with no differences in tumor mutational burden (6.0 vs. 7.6 Mut/Mb, p = 0.6018) and tumor cell PD-L1 expression. FHC-high status identifies NSCLC patients with improved outcomes from pembrolizumab but not chemotherapy, independent of somatic DDR gene status. Prospective studies evaluating FHC alongside germline genetic testing are warranted. Supplementary Information The online version contains supplementary material available at 10.1186/s13045-022-01226-2.


Study design
The main aim of this study was to evaluate the role of FHC in a cohort of patients with metastatic NSCLC with a PD-L1 tumour expression ≥ 50%, treated with first-line pembrolizumab monotherapy [1][2][3][4][5][6][7][8]. Following a data request update, 29 institutions participated to the study and retrospectively included patients treated from January 2017 to May 2020.
In order to assess the potential different impact of FHC depending on the treatment strategy, we PD-L1 expression analysis among the entire population has been already reported [1]. Considering that tumour proportion score (TPS) for PD-L1 expression has been validated with the 22C3 antibody only, we referred to "PD-L1 expression" throughout all the study. All the immunohistochemical (IHC) analyses were preformed locally at each participating institution, using a different antibodies and expression level is reported only as "≥ 50%", and not as a discrete value, only patients with data availability regarding the absolute value of PD-L1 tumour staining have been included in the association analysis between PD-L1 expression and FHC.
To estimate the differential impact of the FHC across the two populations, we evaluated the impact of FHC on clinical outcomes after a perfect random case-control matching between the two cohorts. Cases and controls were randomly paired on the basis of the FHC, age (< 70 vs. ≥ 70 years old), ECOG-PS (0-1 vs ≥ 2), and burden of disease (≥ 2 vs < two metastatic sites).
We then explored the impact of the FHC within the pembrolizumab and chemotherapy cohorts using univariable analyses. A fixed regression model including major determinants of clinical outcome within the study population [1][2][3][4][5][6][7][8] was used for the multivariable analysis of the pembrolizumab cohort.
Additionally, to further evaluate the role of FHC depending on the treatment modality (immunotherapy vs chemotherapy), we performed a pooled analysis of both the cohorts, with and without the interaction term between the FHC and the therapeutic modality (pembrolizumab vs chemotherapy).

Definition of family history of cancer
Family history data was collected from medical records as previously described, and all oncological disease with malignant potential, both hematological and solid, were screened [9]. Lineal line (descendants or ascendants) and collateral line (non-descentants/ascendants e.g., brothers/sisters) were screened till the second degree (grandparents for lineal line and brothers/sisters for the collateral line).
Patients were categorized as follow: FHC-high (in case of at least one cancer diagnosis in both lineal and collateral family lines), FHC-low (in case of at least one cancer diagnosis in either the lineal or collateral line) and FHC-negative (Figure 1). On the basis of our previous findings [9], FHC-high was considered the group of interest for all analyses.

DDR genes exploratory analysis -FDx cohort.
We used a parallel cohort of patients with NSCLC from 4 of the participating institutions (reported in XRCC2). We also explored the associations between FHC, median tumor mutational burden (TMB), smoking status and PD-L1 tumour expression.

Statistical Analysis
The sample size was estimated for the pembrolizumab cohort only, on the basis of the expected number of FHC high patients. According to the subgroup analysis on NSCLC patients evaluated within our previous study [9], we hypothesized a 11% prevalence of FHC high patients and assumed a survival benefit for FHC high patients compared to non-FHC high, with a reduction of the risk of death by 56%.
With a probability of Type I error of 0.05 and of Type II error of 0.20, 238 total events were necessary and at least 633 patients had to be recruited overall from the original cohort. Baseline patients' characteristics were reported with summative descriptive statistics (means, medians and proportions) as appropriate. χ2 test and Fisher's exact test were used to compare categorical variables as appropriate.