Abstract
Purpose:
NRG/RTOG 1016 was a phase III noninferiority trial comparing IMRT + cisplatin versus IMRT + cetuximab for human papillomavirus–positive oropharyngeal squamous cell cancer (HPV+ OPSCC). A germline mutation (the KRAS-variant) previously identified patients with improved outcomes to radiation + cetuximab + cisplatin; thus, we investigated whether there may be similar benefits for IMRT + cetuximab.
Patients and Methods:
562 patients were tested, with 16% positive for the KRAS-variant. Clinical endpoints were per NRG/RTOG 1016, and hazard ratios (HR) were estimated by Cox models. Negative binomial regression modeled treatment-related acute and late grade 3 to 5 adverse events (AE). All models included KRAS, assigned treatment, their interaction, and other prognostic variables tested at two-sided 0.05.
Results:
There was no association between the KRAS-variant and clinical outcomes. There were higher local–regional failure rates (LRF) in nonvariant patients treated with cetuximab [HR = 1.83 (1.19–2.82)] versus KRAS-variant patients [HR (95% CI), 0.90 (0.34–2.41)]. Grade 3 to 4 acute mucositis was higher in nonvariant patients treated with cetuximab [OR = 1.41 (95% CI, 0.98–2.03)] and lower in KRAS-variant patients [OR = 0.60 (95% CI, 0.25–1.41)]. However, the mean late toxicity ratio was lower in nonvariant patients [0.55 (95% CI, 0.35–0.87)] versus KRAS-variant patients [1.62 (95% CI, 0.57–4.62)]. KRAS by treatment interaction for acute and late AEs was not significant (P = 0.0717 and 0.0659, respectively).
Conclusions:
Possible lower LRF and less acute toxicity with cetuximab versus cisplatin for KRAS-variant patients seem to be offset by potentially increased late toxicity. Further evaluation of this class of biomarkers is warranted.
Significance:
Germline microRNA-disrupting biomarkers are functional biomarkers being studied to help guide treatment personalization in cancer. The work in this study is a hypothesis-generating evaluation of the KRAS-variant as a potential biomarker to help predict outcomes and toxicity in patients with head and neck cancer treated with radiation and cetuximab. Findings from this work suggest that these biomarkers may have the potential to differentiate patients at risk of acute versus late toxicity, supporting further study of these biomarkers in radiation studies.
Introduction
There has been a significant effort to improve the local control of tumors treated with radiation, without causing additional toxicity. This has been especially important, yet challenging, in the treatment of head and neck cancer (HNSCC) due to the nature of the areas being treated. The addition of platinum chemotherapy to radiotherapy in HNSCC has been shown to significantly improve survival, however, with additional acute (1) and late toxicity (2). Early efforts to add cetuximab, an antibody targeting the epidermal growth factor receptor (EGFR), to radiotherapy in HNSCC were promising as it was shown to improve survival without an increase in toxicity over radiotherapy alone (3, 4). Based on these data, but without a direct comparison, these two regimens, radiation plus cisplatin, or radiation plus cetuximab, were proposed to have similar efficacy in HNSCC (5). However, the actual risks and benefits of cisplatin versus cetuximab with radiation were unknown.
A significant advance in HNSCC was the identification of a subgroup of patients with HNSCC who were younger and had improved survival—those with oropharyngeal squamous cell carcinomas that were associated with human papillomavirus (HPV) infection (p16-positive; ref. 6). Because of their improved prognosis and younger age, there has been strong interest in de-escalating therapy with the goal of decreasing late toxicity for these patients. NRG/RTOG 1016 was a randomized, noninferiority, multicenter trial for patients with locoregionally advanced p16-positive oropharyngeal cancer (HPV+ OPSCC) to test the hypothesis that treatment with radiation plus cetuximab would result in noninferior survival outcomes with less late toxicity for these patients compared with radiation plus cisplatin (7).
Unfortunately, the results from NRG/RTOG 1016, as well as a similar trial in low-risk HPV+ OPSCC (De-ESCALate HPV), and the TROG 12.01 trial were all negative (7–9). In fact, all trials reported worse cancer outcomes in the cetuximab-treated arms, with worse 2-year overall survival (OS) and 2-year recurrence in low-risk HPV+ OPSCC (De-ESCALaTE HPV; ref. 8), worse progression-free survival (PFS) and locoregional failure (LRF) for all-risk HPV+ OPSCC (RTOG 1016; ref. 7), and worse failure-free survival without improvement in symptom burden or toxicity in low-risk HPV+ OPSCC (TROG 12.01; ref. 9). Interestingly, NRG/RTOG 1016 reported a decrease in acute toxicity and similar late toxicity for cetuximab-treated versus cisplatin-treated patients, and De-ESCALate and TROG found no differences in acute or late toxicity.
A germline microRNA-disrupting variant in the regulatory region of KRAS, referred to as the KRAS-variant, has previously been shown to predict a positive response to cetuximab in retrospective analyses of prospective clinical trials for several cancers, including colon, lung [non–small cell lung cancer (NSCLC)], and HNSCC (10–12). In HNSCC, the KRAS-variant was evaluated in the RTOG 0522 trial, which compared radiation plus cisplatin with or without the addition of cetuximab. The KRAS-variant was found in 17% of RTOG 0522 patients, and these patients were found to benefit significantly from the addition of cetuximab to radiotherapy and cisplatin chemotherapy, with improved PFS for 1 year and OS for 2 years after treatment, which seemed to be most impactful for HPV+ patients. The positive response was hypothesized to be due to the immune-stimulating effects of cetuximab (13–15), as this study also demonstrated that KRAS-variant patients had significantly elevated TGF-B1 compared with nonvariant patients, and thus KRAS-variant patients were likely immunosuppressed (16). KRAS-variant patients were found to have increased grades 3 and 4 mucositis with radiation plus cisplatin, which was not worsened by the addition of cetuximab (47.4% vs. 50%). In contrast, nonvariant patients had lower grades 3 and 4 mucositis with radiation and cisplatin, which was significantly increased with the addition of cetuximab (37.9% vs. 50.6%, P = 0.02). Acute and late toxicity were not separated as distinct endpoints in this analysis.
Because KRAS-variant patients seemed to respond to cetuximab and radiation (in combination with cisplatin) and seemed to have increased radiosensitivity to radiation plus cisplatin that was not worsened by the addition of cetuximab, we evaluated KRAS-variant patients in NRG/RTOG 1016 to see whether there were similar associations with the endpoints of response or toxicity in this study of radiation plus cisplatin versus radiation plus cetuximab.
Patients and Methods
Germline KRAS-variant detection in DNA
DNA from 562 samples was extracted from plasma, buffy coat, or whole blood and was tested for the KRAS-variant using a Taqman-based assay and run in duplicate with appropriate positive and negative controls at MiraDx, a California State Clinical Laboratory Improvement Amendments–certified and College of American Pathologists approved laboratory. A sample was considered positive if it was found to be T/G or G/G, thus heterozygous or homozygous, and wild-type (nonvariant) if T/T.
Statistical methods
Per NRG/RTOG 1016, outcomes were defined as the following: an event for OS is death due to any cause; an event for PFS is local, regional, or distant progression or death due to any cause; an event for LRF is local or regional progression, salvage surgery of the primary tumor with tumor present/unknown, salvage neck dissection with tumor present/unknown >20 weeks after the end of radiotherapy, death due to study cancer without documented progression, or death due to unknown causes without documented progression; distant metastasis (DM) and death from other causes were considered competing risks. LRF and death were considered competing risks for the DM endpoint. A patient could not be considered an event for both LRF and DM endpoints. All endpoints were measured from the date of randomization. OS and PFS and number of grade 3 to 5 adverse events (AE) are protocol-specified analyses, but LRF, DM, and AE rates for specific terms are unplanned. SAS (RRID: SCR_008567) was used for statistical analysis.
Median follow-up was estimated by the reverse Kaplan–Meier method. OS and PFS rates were estimated by the Kaplan–Meier method, and LRF and DM rates by the cumulative incidence method. Hazard ratios (HR) for OS and PFS were estimated by Cox proportional hazards models, and for LRF and DM by cause-specific Cox models. Models were compared by Bayesian information criterion (BIC); lower BIC indicates better fit. The Cox model assumptions were assessed using diagnostics based on scaled Schoenfeld residuals and cumulative martingale residuals. Cox models to assess the prognostic role of the KRAS-variant were stratified by assigned treatment; the KRAS effect was tested at one-sided 0.025 for OS and one-sided 0.05 for PFS, LRF, and DM.
To assess the predictive role of the KRAS-variant, Cox models with KRAS, assigned treatment, and their interaction were used; KRAS by treatment interaction effect was tested at two-sided 0.05 for all endpoints. Three sets of multivariable models were performed for each endpoint: base multivariable model without adjustment for additional covariates, full multivariable model adjusted for all covariates, and reduced multivariable model with minimum BIC of all models evaluated. In addition to KRAS and assigned treatment, the following variables were considered for inclusion in the models: age, gender, Zubrod performance status, smoking history measured in pack-years, T stage, N stage, and risk group per NRG/RTOG 0129.
Rate ratios were estimated using negative binomial regression modeling the number of treatment-related (definitely, probably, or possibly related to treatment) acute (≤180 from end of treatment) and late (>180 from end of treatment) grade 3 to 5 AEs. Common Terminology Criteria for Adverse Events v. 4 was used for toxicity scoring. Negative binomial regression models included KRAS, assigned treatment, and their interaction; the KRAS X treatment interaction effect was tested at two-sided 0.05. Odds ratios (OR) were estimated using logistic regression modeling the probability of specific treatment-related grade 3 to 4 AE terms. Logistic regression models included KRAS, assigned treatment, and their interaction; the KRAS X treatment interaction effect was tested at two-sided 0.05. Grade 3 to 4 treatment-related AE rates over time are estimated in the subgroups defined by KRAS and assigned treatment at the following time points: during treatment; 1 month after treatment (window: −14 to +28 days); 3 months after treatment (window: −28 to +42 days); 6 months after treatment (window: −42 to +56 days); and 1 to 5 years after treatment (window: ±91 days). The analysis population for OS, PFS, LRF, and DM is randomized and eligible patients who have KRAS data. The analysis population for AEs is randomized and eligible patients who started protocol treatment and have KRAS data.
Model assessment for all outcome endpoints did not suggest serious violations of the Cox model assumptions.
NRG/RTOG 1016 study details, patient inclusion, and exclusion
In NRG/RTOG 1016, patients were stratified by T stage (T1–2 vs. T3–4), N stage (N0–2a vs. N2b–3), Zubrod performance status (0 vs. 1), and smoking history (≤10 pack-years vs. >10 pack-years) and randomized (1:1) to IMRT + cisplatin or IMRT + cetuximab.
A CONSORT flow diagram is shown in Supplementary Fig. S1. Of 987 patients, 849 (86%) were randomized, of whom 805 (94.8%) were included in the analysis of the primary endpoint. Of these 805, 733 (91.1%) consented to blood collection. Of these 733, 562 (76.7%) had sample and data available for KRAS and are included in this analysis. The representativeness of the study participants is addressed in Supplementary Table S13.
Results
Prevalence and patient and tumor characteristics
Patient and tumor characteristics were largely similar for patients included in and excluded from analysis (Supplementary Table S1). However, outcomes for patients included in the analysis tended to be worse relative to patients excluded: HRs were 1.43 [95% confidence interval (CI), 1.04–1.96] for OS, 1.22 (95% CI, 0.93–1.60) for PFS, 1.27 (95% CI, 0.87–1.86) for LRF, and 0.83 (95% CI, 0.52–1.33) for DM.
The KRAS-variant was found in 16% of patients, which was close to the projected prevalence (17%) used in the design stage of the study. Patient and tumor characteristics were similar for patients with and without the KRAS-variant and across the treatment arms (Supplementary Table S2). There was a similar percentage of intermediate-risk patients in the KRAS-variant (33.7%) versus nonvariant (29.1%) subgroups. The median follow-up was 8.6 years for the entire cohort, 8.3 years for KRAS-variant patients, and 8.7 years for nonvariant patients.
Outcomes
The KRAS-variant was not a prognostic marker for OS, PFS, LRF, or DM (Supplementary Tables S3–S6; Supplementary Figs. S2–S5), in agreement with previous findings (16). The KRAS-variant was also not a predictive biomarker for OS or PFS, with HRs for treatment effects between variant and nonvariant patients being similar and in the same direction. In the reduced model for OS, the HRs for treatment effect (IMRT + cetuximab/IMRT + cisplatin) were 1.46 (95% CI, 0.65–3.32) for the KRAS-variant subgroup and 1.20 (95% CI, 0.86–1.67) for the nonvariant subgroup (KRAS by treatment interaction P = 0.6540; Supplementary Table S7; Supplementary Fig. S6). For PFS, in the reduced model, the HRs for treatment effect (IMRT + cetuximab/IMRT + cisplatin) were 1.24 (95% CI, 0.61–2.54) in the KRAS-variant subgroup and 1.40 (95% CI, 1.03–1.90) in the nonvariant subgroup (KRAS by treatment interaction P = 0.7700; Supplementary Table S8; Supplementary Fig. S7).
For LRF and DM, although the KRAS-variant was again not predictive, the HRs for treatment effects were in an opposite direction comparing variant and nonvariant patients. For LRF in the base model, the HRs for treatment effect (IMRT + cetuximab/IMRT + cisplatin) were 0.67 (95% CI, 0.25–1.75) for the KRAS-variant subgroup versus 1.67 (95% CI, 1.09–2.57) for the nonvariant subgroup (Fig. 1). In the reduced model, the HRs for treatment effect were 0.90 (95% CI, 0.34–2.41) in the KRAS-variant subgroup and 1.83 (95% CI, 1.19–2.82) in the nonvariant subgroup (KRAS by treatment interaction P = 0.1933; Supplementary Table S9). For DM, in the base model, the HRs for treatment effect (IMRT + cetuximab/IMRT + cisplatin) were 3.19 (95% CI, 0.66–15.34) in the KRAS-variant subgroup and 1.03 (95% CI, 0.57–1.86) in the nonvariant subgroup (Fig. 2). In the reduced model, the HRs for treatment effect are 3.04 (95% CI, 0.63–14.66) in the KRAS-variant subgroup and 1.05 (95% CI, 0.58–1.90) in the nonvariant subgroup (KRAS by treatment interaction P = 0.2160, Supplementary Table S10).
Figure 1.
LRF by KRAS and assigned treatment. Rates of local failure based on treatment and KRAS-variant status, with KRAS-variant patients being depicted by blue lines and nonvariant patients by red lines and dashed lines representing cetuximab-treated patients.
Figure 2.
DM by KRAS and assigned treatment. Distant failure rates based on treatment and KRAS-variant status, with KRAS-variant patients being depicted by blue lines and nonvariant patients by red lines and dashed lines representing cetuximab-treated patients.
Acute toxicity
The mean [standard deviation (SD)] number of acute treatment-related grade 3 to 5 AEs was 3.70 (2.84) for IMRT + cisplatin–treated KRAS-variant patients, 1.98 (2.37) for IMRT + cetuximab–treated KRAS-variant patients, 3.15 (2.95) for IMRT + cisplatin–treated nonvariant patients, and 2.51 (2.40) for IMRT + cetuximab–treated nonvariant patients (Table 1). The mean acute toxicity ratio comparing IMRT + cetuximab/IMRT + cisplatin was 0.53 (95% CI, 0.36–0.80) in the KRAS-variant subgroup and 0.80 (95% CI, 0.67–0.95) in the nonvariant subgroup. This means that the KRAS-variant subgroup of patients treated with IMRT + cetuximab had, on average, 47% fewer acute toxicities compared with KRAS-variant patients treated with IMRT + cisplatin. The nonvariant subgroup of patients treated with IMRT + cetuximab exhibited, on average, 20% fewer acute toxicities compared with those treated with IMRT + cisplatin. However, the interaction of KRAS and treatment was not significant (P = 0.0703).
Table 1.
| KRAS | Assigned treatment | Patients | Mean (SD) | Rate ratio (95% CI) |
|---|---|---|---|---|
| KRAS variant | IMRT + cisplatin | 44 | 3.70 (2.84) | Reference |
| | IMRT + cetuximab | 48 | 1.98 (2.37) | 0.53 (0.36–0.80) |
| Nonvariant | IMRT + cisplatin | 230 | 3.15 (2.95) | Reference |
| | IMRT + cetuximab | 237 | 2.51 (2.40) | 0.80 (0.67–0.95) |
| Total | | 559 | | Interaction P = 0.0703 |
Rate ratios estimated from negative binomial regression model with covariates KRAS (KRAS variant vs. nonvariant), treatment (IMRT + cetuximab vs. IMRT + cisplatin), and the interaction of KRAS and treatment.
Definitely, probably, or possibly related to protocol treatment.
≤180 days from end of treatment.
Scaled Pearson χ2 (value/degrees of freedom) = 0.9754.
Treatment-related grade 3 to 4 acute mucositis, within the KRAS-variant subgroup, was reported in 18 of 44 (40.9%) patients treated with IMRT + cisplatin and in 14 of 48 (29.2%) treated with IMRT + cetuximab. Within the nonvariant subgroup, treatment-related grade 3 to 4 acute mucositis was reported in 94 of 230 (40.9%) patients treated with IMRT + cisplatin and in 117 of 237 (49.4%) treated with IMRT + cetuximab (Table 2). The OR comparing IMRT + cetuximab/IMRT + cisplatin was 0.60 (95% CI, 0.25–1.41) in the KRAS-variant subgroup and 1.41 (95% CI, 0.98–2.03) in the nonvariant subgroup. Again, the interaction of KRAS and treatment was not significant (P = 0.0717). Although not statistically significant, the finding of worse mucositis with the addition of cetuximab to treatment for nonvariant patients, but not for KRAS-variant patients, is consistent with prior findings (16).
Table 2.
| KRAS | Assigned treatment | Patients | Events | OR (95% CI) |
|---|---|---|---|---|
| KRAS variant | IMRT + cisplatin | 44 | 18 (40.9%) | Reference |
| | IMRT + cetuximab | 48 | 14 (29.2%) | 0.60 (0.25–1.41) |
| Nonvariant | IMRT + cisplatin | 230 | 94 (40.9%) | Reference |
| | IMRT + cetuximab | 237 | 117 (49.4%) | 1.41 (0.98–2.03) |
| Total | | 559 | 243 (43.5%) | Interaction P = 0.0717 |
ORs estimated from logistic regression model with covariates KRAS (KRAS variant vs. nonvariant), treatment (IMRT + cetuximab vs. IMRT + cisplatin), and the interaction of KRAS and treatment.
Definitely, probably, or possibly related to protocol treatment.
≤180 days from end of treatment.
Common Terminology Criteria for Adverse Events version 4 term: mucositis oral.
There were no significant differences between KRAS-variant patients versus nonvariant patients for skin reaction inside or outside of the portal (Supplementary Tables S11 and S12).
Late toxicity
The mean (SD) number of late treatment-related grade 3 to 5 AEs was 0.27 (0.55) for IMRT + cisplatin–treated KRAS-variant patients, 0.43 (0.83) for IMRT + cetuximab–treated KRAS-variant patients, 0.49 (1.18) for IMRT + cisplatin–treated nonvariant patients, and 0.27 (0.70) for IMRT + cetuximab–treated nonvariant patients (Table 3). The mean late toxicity ratio comparing IMRT + cetuximab/IMRT + cisplatin was 1.62 (95% CI, 0.57–4.62) in the KRAS-variant subgroup and 0.55 (95% CI, 0.35–0.87) in the nonvariant subgroup. This means that the KRAS-variant subgroup of patients treated with IMRT + cetuximab exhibited, on average, 62% more late toxicities compared with those treated with IMRT + cisplatin. The nonvariant subgroup of patients treated with IMRT + cetuximab exhibited, on average, 45% fewer late toxicities compared with those treated with IMRT + cisplatin. The interaction of KRAS and treatment was again not significant (P = 0.0659).
Table 3.
| KRAS | Assigned treatment | Patients | Mean (SD) | Rate ratio (95% CI) |
|---|---|---|---|---|
| KRAS variant | IMRT + cisplatin | 41 | 0.27 (0.55) | Reference |
| | IMRT + cetuximab | 46 | 0.43 (0.83) | 1.62 (0.57–4.62) |
| Nonvariant | IMRT + cisplatin | 221 | 0.49 (1.18) | Reference |
| | IMRT + cetuximab | 225 | 0.27 (0.70) | 0.55 (0.35–0.87) |
| Total | | 533 | | Interaction P = 0.0659 |
Rate ratios estimated from negative binomial regression model with covariates KRAS (KRAS variant vs. nonvariant), treatment (IMRT + cetuximab vs. IMRT + cisplatin), and the interaction of KRAS and treatment.
Definitely, probably, or possibly related to protocol treatment.
>180 days from end of treatment.
Scaled Pearson χ2 (value/degrees of freedom) = 0.9778.
Grade 3 to 4 treatment-related AE rates over time are shown in Fig. 3, allowing the view of toxicity across both subgroups (KRAS-variant vs. nonvariant) and all treatment approaches (IMRT + cisplatin vs. IMRT + cetuximab). Visually, this seems to show lower acute toxicity for KRAS-variant patients treated with cetuximab compared with KRAS-variant patients treated with cisplatin, and nonvariant patients treated with cetuximab or cisplatin through 6 months, but then loss of these differences by 1 year.
Figure 3.
Grade 3 to 4 treatment-related AEs over time by KRAS and assigned treatment. Rates of treatment-related AEs with 95% CI by treatment and KRAS-variant status, with KRAS-variant patients being depicted by blue (cisplatin-treated) and red (cetuximab-treated) and nonvariant patients depicted by green (cisplatin-treated) and brown (cetuximab-treated).
Discussion
In this study, we tested the hypothesis that there would be an association of the KRAS-variant, the first germline microRNA-associated mutation discovered in cancer, and HPV+ OPSCC patient outcome with treatment with radiation plus cetuximab versus radiation plus cisplatin. This study did not support the hypothesis that the KRAS-variant is a predictive biomarker of improved outcome in patients with HPV+ OPSCC treated with IMRT + cetuximab. However, there was a nonsignificant trend for a lower rate of LRF with cetuximab for the KRAS-variant subgroup of patients, in contrast to what is seen in nonvariant patients who exhibit a significantly higher risk of LRF. In addition, acute toxicity for KRAS-variant patients seemed to be nonsignificantly lower with cetuximab treatment than all other treatments for either subgroup. Unfortunately however, late toxicity for KRAS-variant patients treated with cetuximab was nonsignificantly worse than for cisplatin-treated KRAS-variant patients.
There have been numerous studies on the role of cetuximab in patients with the KRAS-variant. Surprisingly, although tumor-acquired KRAS mutations predict a poor response to cetuximab, in patients with colon cancer, KRAS-variant patients seem to respond to single-agent cetuximab, even in the context of a tumor KRAS mutation, without any additional benefit from chemotherapy (10). Importantly, the role of cetuximab in the treatment of KRAS-variant patients is strongly affected by how cetuximab is given, such as alone or in combination with other cytotoxic agents. For example, in metastatic colon cancer, cetuximab treatment alone improves survival significantly for these patients (10, 17, 18), but if given in combination with irinotecan, KRAS-variant patients had a significantly worse outcome (19).
Two studies have investigated the role of cetuximab in combination with radiation and chemotherapy for KRAS-variant patients. The NRG/RTOG 0522 trial included cisplatin with radiation with or without cetuximab in HNSCC (16), and the NRG/RTOG 0617 trial in NSCLC(12) included combination chemotherapy with radiation with or without cetuximab. Both studies suggested that there may be a benefit in outcome for KRAS-variant patients with the addition of cetuximab to chemotherapy and radiation. Of note, both of these prior studies also suggested that KRAS-variant patients had increased radiosensitivity with chemotherapy and radiation, which in NRG/RTOG 0522 was not worsened with the addition of cetuximab (16) and in NRG/RTOG 0617 was worsened with the addition of cetuximab (12).
This study is the first evaluation of radiation with cetuximab alone in KRAS-variant patients. In addition, this study is the only study thus far to evaluate the impact of cetuximab on acute versus late radiation toxicity for KRAS-variant patients. This may be an important distinction, as we found here that acute toxicity with cetuximab and radiation versus cisplatin and radiation was numerically, although not significantly, lower for KRAS-variant patients. In addition, acute toxicity with cetuximab was lower for KRAS-variant patients compared to nonvariant patients treated with either cisplatin or cetuximab. However, there seemed to be higher late toxicity for IMRT + cetuximab–treated KRAS-variant patients compared with cisplatin-treated KRAS-variant patients. This increased late toxicity for KRAS-variant patients could be explained by the significant higher circulating TGFB1 levels previously shown in KRAS-variant patients with HNSCC (16), as TGFB1 has been shown to be an essential factor in the development of late toxicity (20).
The different predictive power of the KRAS-variant for acute versus late toxicity is both interesting and could be important. In fact, there are now numerous studies showing that the class of germline microRNA-associated mutations (mirSNPs), of which the KRAS-variant was the first example, contains powerful biomarkers identifying patients who have different toxicity risks to radiotherapy. The first study showing this was in sarcoma, in which a mirSNP signature identified patients at increased risk of major wound toxicity after steriotactic body radiation therapy (SBRT; ref. 21). Subsequent studies have focused on prostate cancer, in which mirSNP signatures have been found to predict late toxicity to both SBRT and conventionally fractionated radiation therapy CFRT, and the mirSNP signatures differ between the two treatment approaches (22, 23). A recent study also found a mirSNP signature predicting acute toxicity to prostate cancer SBRT that is unique from the signature of late toxicity (Weidhaas, 2025), which further validates the unique genetic responses leading to these two toxicity endpoints. This is consistent with our preliminary findings in this study, showing that KRAS-variant patients could experience less acute toxicity but more late toxicity.
Although major limitations of the study include the small number of KRAS-variant patients, significantly limiting statistical power, this study is the largest study of HPV+ OPSCC that exists. It is also very likely based on additional mirSNP studies discussed above that radiation toxicity prediction requires more than a single mirSNP, and only the KRAS-variant was evaluated in this study. Regardless, although this study does not support the use of the KRAS-variant alone in directing cetuximab therapy for patients with HPV+ OPSCC, this study does expand the growing evidence supporting the potential of this class of biomarkers in the personalization of radiotherapy. With patient-specific genetic mirSNP information, there is hope that in the near future we will be able to not only consider patient response to treatment but also their individual acute and late toxicity risks. These are critical steps toward comprehensive treatment personalization for patients being treated with radiotherapy.
Supplementary Material
Missing Data Analysis
CONSORT flow diagram
Patient and Tumor Characteristics by KRAS and Assigned Treatment
Overall Survival by KRAS
Univariate and Multivariable Cox Models for KRAS as a Prognostic Biomarker for Overall Survival
Progression-Free Survival by KRAS
Univariate and Multivariable Cox Models for KRAS as a Prognostic Biomarker for Progression-Free Survival
Local-Regional Failure by KRAS
Univariate and Multivariable Cause-Specific Cox Models for KRAS as a Prognostic Biomarker for Local-Regional Failure
Distant Metastasis by KRAS
Univariate and Multivariable Cause-Specific Cox Models for KRAS as a Prognostic Biomarker for Distant Metastasis
Overall Survival by KRAS and Assigned Treatment
Multivariable Cox Models for KRAS as a Predictive Biomarker for Overall Survival
Progression-Free Survival by KRAS and Assigned Treatment
Multivariable Cox Models for KRAS as a Predictive Biomarker for Progression-Free Survival
Multivariable Cox Models for KRAS as a Predictive Biomarker for Local-Regional Failure
Multivariable Cox Models for KRAS as a Predictive Biomarker for Distant Metastasis
Grade 3-4 Treatment-Related [1] Skin Reaction Inside Portal [2] by KRAS and Assigned Treatment
Grade 3-4 Treatment-Related [1] Skin Reaction Outside Portal [2] by KRAS and Assigned Treatment
Representativeness of Study Participants
Acknowledgments
This project was supported by grants UG1CA189867 (NCORP), U10CA180868 (NRG Oncology Operations), U10CA180822 (NRG Oncology SDMC), U24CA196067 (NRG Specimen Bank), U24CA180803 (IROC) from the National Cancer Institute (NCI), and Eli Lilly. J.B. Weidhaas was supported by a grant from the NCI (R01CA 238998).
Footnotes
Note: Supplementary data for this article are available at Cancer Research Communications Online (https://aacrjournals.org/cancerrescommun/).
Data Availability
The data generated in this study are not publicly available as patients did not consent to individual genetic data being shared, but group data are available upon reasonable request to the corresponding author.
Authors’ Disclosures
J.B. Weidhaas reports a patent to “The KRAS variant in cancer” issued and licensed and has co-founded a company, MiraDx, which has licensed a patent from Yale University about the KRAS-variant. K. Higgins reports personal fees from AstraZeneca, and Daiichi Sankyo and other support from Picture Health during the conduct of the study. J.J. Caudell reports grants from Varian Medical Systems outside the submitted work. S.S. Yom reports grants from Merck, Bristol Myers Squibb, EMD Serono, and Johnson & Johnson outside the submitted work. No disclosures were reported by the other authors.
Authors’ Contributions
J.B. Weidhaas: Conceptualization, investigation, methodology, writing–original draft. J. Harris: Resources, data curation, software, formal analysis, visualization, methodology, writing–review and editing. M.L. Gillison: Resources, software, methodology, writing–review and editing. D. Blakaj: Resources. G.A. Krempl: Investigation, writing–review and editing. K. Higgins: Resources, writing–original draft. J. Phan: Data curation, writing–original draft, writing–review and editing. N.E. Dunlap: Conceptualization, writing–review and editing. S.T. Mahamood: Resources. J. Dorth: Resources, writing–review and editing. J.J. Caudell: Resources, writing–review and editing. A.B. Desai: Resources, writing–review and editing. T.J. Galloway: Writing–review and editing. J.D. Pennington: Resources, writing–review and editing. A. Currey: Investigation, writing–review and editing. J. Lathrop: Investigation. P.A. Torres-Saavedra: Formal analysis, visualization, methodology, writing–original draft, writing–review and editing. D.N. Hayes: Formal analysis, validation, visualization, project administration, writing–review and editing. S.S. Yom: Resources, supervision, visualization, project administration, writing–review and editing. Q.-T. Le: Conceptualization, writing–review and editing.
References
- 1. Adelstein DJ, Li Y, Adams GL, Wagner H Jr, Kish JA, Ensley JF, et al. An intergroup phase III comparison of standard radiation therapy and two schedules of concurrent chemoradiotherapy in patients with unresectable squamous cell head and neck cancer. J Clin Oncol 2003;21:92–8. [DOI] [PubMed] [Google Scholar]
- 2. Machtay M, Moughan J, Trotti A, Garden AS, Weber RS, Cooper JS, et al. Factors associated with severe late toxicity after concurrent chemoradiation for locally advanced head and neck cancer: an RTOG analysis. J Clin Oncol 2008;26:3582–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Bonner JA, Harari PM, Giralt J, Azarnia N, Shin DM, Cohen RB, et al. Radiotherapy plus cetuximab for squamous-cell carcinoma of the head and neck. N Engl J Med 2006;354:567–78. [DOI] [PubMed] [Google Scholar]
- 4. Bonner JA, Harari PM, Giralt J, Cohen RB, Jones CU, Sur RK, et al. Radiotherapy plus cetuximab for locoregionally advanced head and neck cancer: 5-year survival data from a phase 3 randomised trial, and relation between cetuximab-induced rash and survival. Lancet Oncol 2010;11:21–8. [DOI] [PubMed] [Google Scholar]
- 5. Levy AR, Johnston KM, Sambrook J, Donato B, Penrod JR, Corral M, et al. Indirect comparison of the efficacy of cetuximab and cisplatin in squamous cell carcinoma of the head and neck. Curr Med Res Opin 2011;27:2253–9. [DOI] [PubMed] [Google Scholar]
- 6. Ang KK, Harris J, Wheeler R, Weber R, Rosenthal DI, Nguyen-Tân PF, et al. Human papillomavirus and survival of patients with oropharyngeal cancer. N Engl J Med 2010;363:24–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Gillison ML, Trotti AM, Harris J, Eisbruch A, Harari PM, Adelstein DJ, et al. Radiotherapy plus cetuximab or cisplatin in human papillomavirus-positive oropharyngeal cancer (NRG Oncology RTOG 1016): a randomised, multicentre, non-inferiority trial. Lancet 2019;393:40–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Mehanna H, Robinson M, Hartley A, Kong A, Foran B, Fulton-Lieuw T, et al. Radiotherapy plus cisplatin or cetuximab in low-risk human papillomavirus-positive oropharyngeal cancer (De-ESCALaTE HPV): an open-label randomised controlled phase 3 trial. Lancet 2019;393:51–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Rischin D, King M, Kenny L, Porceddu S, Wratten C, Macann A, et al. Randomized trial of radiation therapy with weekly cisplatin or cetuximab in low-risk HPV-associated oropharyngeal cancer (TROG 12.01) - a Trans-Tasman Radiation Oncology Group Study. Int J Radiat Oncol Biol Phys 2021;111:876–86. [DOI] [PubMed] [Google Scholar]
- 10. Saridaki Z, Weidhaas J, Lenz H-J, Laurent-Puig P, Jacobs B, De Schutter J, et al. A let-7 microRNA-binding site polymorphism in KRAS predicts improved outcome in metastatic colorectal cancer (mCRC) patients treated with salvage cetuximab/panitumumab monotherapy. Clin Cancer Res 2014;20:4499–510. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Chung CH, Lee JW, Slebos RJ, Howard JD, Perez J, Kang H, et al. A 3′-UTR KRAS-variant is associated with cisplatin resistance in patients with recurrent and/or metastatic head and neck squamous cell carcinoma. Ann Oncol 2014;25:2230–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Weidhaas JB, Hu C, Komaki R, Masters GA, Blumenschein GR, Chang JY, et al. The inherited KRAS-variant as a biomarker of cetuximab response in NSCLC. Cancer Res Commun 2023;3:2074–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Yang X, Zhang X, Mortenson E, Radkevich-Brown O, Wang Y, Fu Y-X. Cetuximab-mediated tumor regression depends on innate and adaptive immune responses. Mol Ther 2013;21:91–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Srivastava R, Lee S, Andrade Filho P, Lord C, Jie H, Davidson H, et al. Cetuximab-activated natural killer and dendritic cells collaborate to trigger tumor antigen-specific T-cell immunity in head and neck cancer patients. Clin Cancer Res 2013;19:1858–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Stephenson R, Lim C, Matthews M, Dietsch G, Hershberg R, Ferris R. TLR8 stimulation enhances cetuximab-mediated natural killer cell lysis of head and neck cancer cells and dendritic cell cross priming of EGFR-specific CD8+ T cells. Cancer Immunol Immunother 2013;62:1347–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Weidhaas JB, Harris J, Schaue D, Chen AM, Chin R, Axelrod R, et al. The KRAS-variant and cetuximab response in head and neck squamous cell cancer: a secondary analysis of a randomized clinical trial. JAMA Oncol 2017;3:483–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Zhang W, Winder T, Ning Y, Pohl A, Yang D, Kahn M, et al. A let-7 microRNA-binding site polymorphism in 3′-untranslated region of KRAS gene predicts response in wild-type KRAS patients with metastatic colorectal cancer treated with cetuximab monotherapy. Ann Oncol 2011;22:104–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Zhang W, Labonte MJ, Lenz HJ. KRAS let-7 LCS6 SNP predicts cetuximab efficacy in KRASwt metastatic colorectal cancer patients: does treatment combination partner matter? Ann Oncol 2011;22:484–5. [DOI] [PubMed] [Google Scholar]
- 19. Graziano F, Canestrari E, Loupakis F, Ruzzo A, Galluccio N, Santini D, et al. Genetic modulation of the Let-7 microRNA binding to KRAS 3′-untranslated region and survival of metastatic colorectal cancer patients treated with salvage cetuximab-irinotecan. Pharmacogenomics J 2010;10:458–64. [DOI] [PubMed] [Google Scholar]
- 20. Martin M, Lefaix J, Delanian S. TGF-beta1 and radiation fibrosis: a master switch and a specific therapeutic target? Int J Radiat Oncol Biol Phys 2000;47:277–90. [DOI] [PubMed] [Google Scholar]
- 21. Kalbasi A, Kamrava M, Chu F-I, Telesca D, Van Dams R, Yang Y, et al. A phase 2 trial of five-day neoadjuvant radiation therapy for patients with high-risk primary soft tissue sarcoma. Clin Cancer Res 2020;3524:2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Kishan AU, Marco N, Schulz-Jaavall M-B, Steinberg ML, Tran PT, Juarez JE, et al. Germline variants disrupting microRNAs predict long-term genitourinary toxicity after prostate cancer radiation. Radiother Oncol 2022;167:226–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Kishan AU, Marco N, Ma TM, Steinberg ML, Sachdeva A, Cao M, et al. Application of a genetic signature of late GU toxicity in SCIMITAR, a Post-op SBRT trial. Clin Transl Radiat Oncol 2023;39:100594. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Missing Data Analysis
CONSORT flow diagram
Patient and Tumor Characteristics by KRAS and Assigned Treatment
Overall Survival by KRAS
Univariate and Multivariable Cox Models for KRAS as a Prognostic Biomarker for Overall Survival
Progression-Free Survival by KRAS
Univariate and Multivariable Cox Models for KRAS as a Prognostic Biomarker for Progression-Free Survival
Local-Regional Failure by KRAS
Univariate and Multivariable Cause-Specific Cox Models for KRAS as a Prognostic Biomarker for Local-Regional Failure
Distant Metastasis by KRAS
Univariate and Multivariable Cause-Specific Cox Models for KRAS as a Prognostic Biomarker for Distant Metastasis
Overall Survival by KRAS and Assigned Treatment
Multivariable Cox Models for KRAS as a Predictive Biomarker for Overall Survival
Progression-Free Survival by KRAS and Assigned Treatment
Multivariable Cox Models for KRAS as a Predictive Biomarker for Progression-Free Survival
Multivariable Cox Models for KRAS as a Predictive Biomarker for Local-Regional Failure
Multivariable Cox Models for KRAS as a Predictive Biomarker for Distant Metastasis
Grade 3-4 Treatment-Related [1] Skin Reaction Inside Portal [2] by KRAS and Assigned Treatment
Grade 3-4 Treatment-Related [1] Skin Reaction Outside Portal [2] by KRAS and Assigned Treatment
Representativeness of Study Participants
Data Availability Statement
The data generated in this study are not publicly available as patients did not consent to individual genetic data being shared, but group data are available upon reasonable request to the corresponding author.