Oncology · Medical Literature
Afatinib for Advanced Cancers Carrying an EGFR, HER2, or HER3 Mutation: An Open-Label, Phase II, Belgian Precision Study.
Abstract
The Belgian Precision initiative aims to implement tumor-agnostic next-generation sequencing (NGS) in patients with advanced cancer and expand genotype-matching drug availability. The present investigator-driven trial aimed to study the efficacy of afatinib in patients with advanced solid tumors harboring an activating HER2, EGFR, or HER3 mutation. This open-label phase II trial has three cohorts: HER2-, EGFR-, and HER3-mutated previously treated solid tumors. The primary end point was objective response rate (ORR). For each cohort, a Simon two-stage design was used. Observation of ≥2 responses in the first 10 patients prompted a further 19 to be included. Secondary end points were disease control rate (DCR), duration of response (DOR), progression-free survival, overall survival, and safety. A total of 45 patients were included, with a median age of 62 years. For the HER2 cohort (n = 30), ORR was 3.3% (one partial response) and DCR was 23.3%. In the EGFR cohort, a total of seven patients were included, resulting in an ORR in 2/7 (28.6%) patients, with a DOR of 6.6 and 15.4 months. In the HER3 cohort, a total of eight patients were included, but none demonstrated an objective response. Safety data were consistent with the known safety profile of afatinib. The present phase II study, investigating afatinib in HER2-, EGFR-, orHER3-mutant pretreated solid tumors did not reach its primary end point in the HER2 cohort. Neither the EGFR nor the HER3 cohort reached full accrual, but clinically meaningful responses were observed in two EGFR-mutated patients. Further exploration of HER targeting in solid tumors is warranted.
Clinical relevance
In previously treated patients with advanced solid tumors harboring an activating EGFR, HER2, or HER3 mutation (identified by tumor-agnostic NGS), does afatinib monotherapy achieve a clinically meaningful objective response rate (ORR) across three genotype-defined cohorts? __SNAPSHOT__ {"studyDesign":"Investigator-driven, open-label, multicenter, phase II basket trial with three parallel genotype-defined cohorts (HER2-mutated, EGFR-mutated, HER3-mutated); Simon two-stage design per cohort","setting":"Belgium; multicenter academic setting (Belgian Precision / PRECISION 2 national initiative); NCT03810872","sampleSize":"n=45 total (HER2 cohort: n=30; EGFR cohort: n=7; HER3 cohort: n=8); only the HER2 cohort completed full planned accrual","population":"Adults with locally advanced or metastatic solid tumors harboring an activating EGFR (non-NSCLC), HER2, or HER3 somatic mutation identified by NGS; failure of at least one prior line of standard systemic therapy; ECOG PS ≤2; adequate organ function. Median age 62 years. Tumor types in HER2 cohort included NSCLC (n=11), breast (n=8), cervix (n=3), colorectal (n=2), head and neck (n=2), and others. EGFR cohort included pancreas, head and neck, breast, endocrine tumor, cervix, and kidney. HER3 cohort included breast, pancreas, colorectal, head and neck, esophagus, and stomach cancers.","intervention":"Afatinib monotherapy, oral, 40 mg once daily in 28-day cycles (second-generation, irreversible pan-HER TKI inhibiting EGFR, HER2, HER4, and blocking HER3 transphosphorylation)","comparator":"Not applicable (single-arm study; no control arm)","primaryOutcome":"Objective response rate (ORR) per cohort, assessed by RECIST 1.1 every 8 weeks. Simon two-stage design tested null hypothesis ORR=10% vs. alternative ORR=30% (one-sided alpha=0.05, power=80%). Advancement to stage 2 required ≥2 responses in first 10 patients.","followUp":"Not reported as a fixed follow-up duration; duration of response reported for individual responders (6.6 months and 15.4 months in EGFR cohort; 8.5 months in HER2 cohort responder)","fundingSource":"Investigator-driven; partially supported by Kom op tegen Kanker (Belgian cancer charity); industry sponsorship not reported in abstract"}
Key findings
HER2 cohort (n=30; only cohort to complete full accrual): ORR=3.3% (1/30; one partial response in a HER2-mutated NSCLC patient, DOR 8.5 months); DCR=23.3%. The Simon two-stage primary endpoint was not met—fewer than 2 responses were observed among the first 10 patients, but the cohort was nevertheless expanded (full accrual details not further specified in abstract). The ORR of 3.3% falls well below both the null hypothesis threshold of 10% and the target of 30%, indicating no meaningful single-agent activity in this unselected HER2-mutated pan-tumor population. EGFR cohort (n=7; did not reach full accrual of 29): ORR=28.6% (2/7; partial responses in one EGFR-mutated breast cancer patient, DOR 6.6 months, and one EGFR-mutated pancreatic adenocarcinoma patient, DOR 15.4 months); DCR=42.9%. Although the point estimate meets the alternative hypothesis threshold, the extremely small sample (n=7) precludes any definitive conclusion; confidence intervals are not reported. HER3 cohort (n=8; did not reach full accrual): ORR=0% (0/8); DCR=62.5%. No objective responses were observed. Safety: consistent with the established afatinib safety profile (diarrhea, acneiform rash, mucositis); no novel safety signals reported in the abstract. NNT: not calculable given single-arm design. No p-values or confidence intervals are reported in the available abstract.
Practice implications
This study established that afatinib monotherapy does not produce clinically meaningful ORR in the population of patients with previously treated HER2-mutated advanced solid tumors as a genotype-unselected basket: only 1 of 30 patients responded, and the primary endpoint was not met. The study did not establish efficacy in the HER3-mutated cohort (0/8 responses). For the EGFR-mutated non-NSCLC cohort, two clinically durable responses were observed (DOR 6.6 and 15.4 months), suggesting potential activity in specific EGFR-mutated tumor types (notably pancreatic adenocarcinoma and breast cancer), consistent with prior case reports and the NCI-MATCH experience; however, with only 7 patients enrolled and no confidence intervals reported, this signal is hypothesis-generating only and cannot support practice change. Importantly, the EGFR and HER3 cohorts did not reach planned accrual, substantially limiting interpretability. The heterogeneous tumor types within each cohort preclude tumor-type-specific conclusions. Patients with HER2-amplified (non-mutated) tumors, EGFR-mutated NSCLC, ECOG PS >2, and patients who were anti-HER2 therapy-naïve are not well represented and results should not be extrapolated to those populations. The lack of specific mutation subtype characterization in the abstract is a notable gap, as mutation-specific afatinib sensitivity is likely. Only the abstract is available; full PFS, OS, and detailed safety data are not assessable from this summary. Practice should not change based on these results. NOTE: This summary is based on the published abstract only; full manuscript data may contain additional findings not captured here. __LIMITATIONS__ Only the HER2 cohort completed planned accrual (n=30); both the EGFR (n=7/29 planned) and HER3 (n=8/29 planned) cohorts were significantly underpowered, likely due to challenges in systematic tumor-agnostic NGS implementation and COVID-19 disruptions to recruitment.,Extremely small sample sizes in the EGFR (n=7) and HER3 (n=8) cohorts render all point estimates highly unstable; confidence intervals are not reported, making it impossible to rule out chance findings.,Basket design with marked tumor-type heterogeneity within each cohort: HER2 cohort alone included 9 different tumor types, making genotype-by-histology interaction effects undetectable.,No data on specific mutation subtypes or alleles within each cohort, which is critical given known mutation-specific differences in afatinib sensitivity.,Single-arm design with no comparator arm; no randomization; selection bias cannot be excluded.,No blinding (open-label); though RECIST is an objective endpoint, investigator bias in response assessment cannot be entirely excluded.,PFS, OS, and detailed safety data not reported in the abstract; full clinical benefit profile cannot be assessed.,No confidence intervals or p-values reported in the abstract for any outcome.,Generalizability is limited to Belgium's patient population accessing NGS-guided enrollment; systematic NGS access was inconsistent, as acknowledged by the authors.,Statistical significance of the EGFR cohort ORR (28.6% in n=7) cannot be evaluated; the Simon two-stage threshold was not formally reached due to incomplete accrual.,Industry sponsorship status is not clearly reported; investigator-driven design may reduce but does not eliminate potential conflicts of interest. __BOTTOM_LINE_CATEGORY__ not_ready_for_clinical_application __BOTTOM_LINE__ Afatinib monotherapy does not produce meaningful anti-tumor activity in unselected patients with previously treated HER2-mutated advanced solid tumors (ORR 3.3%), and no responses were observed in the small HER3-mutated cohort. Two durable partial responses in the EGFR-mutated non-NSCLC cohort (ORR 28.6%; n=7) are clinically intriguing—particularly in pancreatic adenocarcinoma and breast cancer—but the cohort was too small and underpowered to support any practice change. These findings do not support routine use of afatinib in HER2- or HER3-mutated solid tumors, and further tumor-type-specific studies with larger, mutation-subtype-selected populations are needed before any clinical adoption in the EGFR-mutated non-NSCLC setting.