A Prospective Study of Definitive ChemoRadiation With ImmunOtherapy (CRIO) in Locally or Regionally Advanced, Unresectable Cutaneous Squamous Cell Carcinoma

Abstract

Purpose

There are limited prospective data to guide the management of patients with cutaneous squamous cell carcinoma (cSCC) presenting with locally advanced, unresectable disease. Previously, we conducted a prospective pilot study of chemoradiation for patients with unresectable cSCC. The purpose of this study was to evaluate the efficacy and safety of chemoradiation combined with checkpoint inhibitor immunotherapy with durvalumab for this patient cohort.

Methods and Materials

Patients with locally advanced and/or nodal metastatic cSCC deemed unsuitable for surgery were treated with definitive radiation therapy to a dose of 70 Gy in 35 fractions, concurrent weekly chemotherapy (cisplatin 40 mg/m² or carboplatin area under curve 2 mg/mL/min), and durvalumab 1500 mg every 4 weeks. 18F-fluorodeoxyglucose positron emission tomography assessment was performed at 19 weeks. Patients who achieved a complete response (CR) did not receive further treatment. Patients with residual, but not progressive, disease could continue durvalumab for a further 6 cycles, with or without salvage surgery. Primary endpoints were CR and safety.

Results

Fifteen patients were enrolled between May 2019 and June 2021. Fourteen (93%) had regional nodal disease classified as nonmetastatic stage IV. CR was observed in 9/15 (60%), with no patients undergoing salvage surgery. With a median follow-up of 39 months, all patients who achieved CR were alive without progression or died from another cause. The 6 patients who did not achieve CR all progressed: 5 died due to disease progression and 1 received further systemic therapy. Grade 3 adverse events related to immunotherapy were reported in 1 patient (7%). The estimated rate of both overall and progression-free survival at 24 months was 65% (95% CI, 35%-84%).

Conclusions

To our knowledge, this is the only prospective series of chemoradiation in combination with checkpoint inhibitor immunotherapy for unresectable cSCC. A high CR rate was achieved, and responses were durable. The number of serious adverse events was limited.

Introduction

Cutaneous squamous cell carcinoma (cSCC) is the second most common skin cancer behind basal cell carcinoma. Worldwide incidence has been increasing,¹ possibly due to an aging population, depletion of the ozone layer, occupational exposure, diet and lifestyle changes, and increased use of immunosuppression. Australia has the highest incidence of nonmelanoma cutaneous malignancies,² and it is estimated that 2 in 3 Australians will develop a skin cancer in their lifetime. The head and neck region accounts for approximately 85% of cutaneous malignancies due to sun exposure. Local excision is the mainstay of treatment for early disease.

Metastatic disease is uncommon (<5%), and often nodal,^3,4 although hospital-referred patients likely have a higher incidence of locally advanced disease and nodal metastases. Surgery with adjuvant radiation therapy is considered standard treatment for advanced resectable disease.⁵ However, surgery may not be appropriate for a proportion of patients due to surgically unresectable disease, medically unsuitable for surgery, or unacceptable functional/cosmetic outcomes. There is no standard of care for these patients, and the chance of cure with definitive radiation therapy alone is considered low.^6,7

Genomic studies have found that cSCCs are characterized by a very high mutational burden,^8,9 comparable to that of head and neck squamous cell carcinoma (SCC).¹⁰ This provides a rationale to employ concurrent DNA-damaging therapies with platinum-based chemotherapy and radiation therapy. Additionally, a high mutational burden and disease incidence in patients receiving immunosuppression suggest a likely response to systemic checkpoint inhibitor therapy.

Prospective phase 2 studies have shown impressive responses with the programmed death 1 (PD-1) inhibitors cemiplimab and pembrolizumab as monotherapy for advanced cSCC.11, 12, 13 In these trials, patients with unresectable locally advanced disease had either previously received, or were unsuitable for, definitive radiation therapy.

Our institution previously published a prospective study of concurrent platinum-based chemotherapy with radiation therapy for patients with unresectable cSCC, demonstrating a complete response (CR) rate of 50% with feasible administration and acceptable toxicity.¹⁴ This prospective single-arm study aimed to document the efficacy and safety of a combined approach of concurrent chemoradiation with the addition of durvalumab, a programmed death ligand 1 (PD-L1) inhibitor (the ChemoRadiation with ImmunOtherapy [CRIO] study).¹⁵

Methods and Materials

Patients

Eligible patients were at least 18 years of age, with histologically confirmed, assessable, unresectable cSCC with locally advanced primary disease or regional nodal metastases. Baseline assessment was performed clinically, radiologically with computed tomography (CT) and/or magnetic resonance imaging, and metabolically with 18F-fluorodeoxyglucose (FDG) positron emission tomography (PET). Patients had an Eastern Cooperative Oncology Group performance status of 1 or less and adequate organ function. Patients with a history of autoimmune disease, immunosuppression, or prior treatment with PD-1/PD-L1 checkpoint inhibitor therapy were excluded.¹⁵

The disease had to be deemed unresectable by the local head and neck and/or complex skin multidisciplinary clinic. Factors used to determine unresectable disease included technical surgical and anatomic considerations, medical unsuitability for surgery, and unacceptability of functional and cosmetic outcomes.

Oversight

The trial was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. Astra Zeneca provided funding for the study and durvalumab but did not contribute to data collection, analysis, or interpretation. No external writers were involved in preparing the manuscript. The trial protocol was granted ethical approval by the local Human Research Ethics Committee. All patients signed informed consent. The trial was registered with the Australian and New Zealand Clinical Trials Registry (ACTRN12618001573246).

Treatment regimen

Radiation therapy was delivered by intensity modulated radiation therapy. Gross primary and nodal disease were treated to 70 Gy in 35 fractions over 7 weeks. Elective nodal irradiation was delivered to the macroscopically occult, first drainage echelon to a total dose of 54 Gy in 35 fractions over 7 weeks, simultaneously integrated. Head and neck lymph node levels were as defined by Grégoire et al.¹⁶

Chemotherapy was to be delivered weekly, intravenously, concurrently with radiation therapy. The preferred regimen was with cisplatin at 40 mg/m² of body surface area. Carboplatin was an alternative for patients with renal, cardiac, or hearing impairment, at the discretion of the treating physician, dosed at the area under the concentration-time curve, 2 mg/mL/min.

Immunotherapy with durvalumab was delivered intravenously at a fixed dose of 1500 mg every 4 weeks, starting at week 1 of chemoradiation. At week 19, after 5 cycles of durvalumab and 12 weeks from completion of radiation therapy, patients were reassessed with CT and/or magnetic resonance imaging, and FDG PET. Patients assessed to have a CR did not receive further treatment. Patients with residual, but not progressive, disease were to be assessed for consideration of salvage surgery. Up to 6 further cycles of durvalumab could be delivered, either alone or as adjuvant treatment to salvage surgery. Treatment was discontinued if unequivocal progression, death, unacceptable toxic effects, or withdrawal of consent occurred.

Endpoints and assessments

All patients underwent clinical, radiological, and metabolic assessment at baseline and at 19 weeks. Where results were equivocal, suspected residual disease was biopsied, or imaging was repeated at 4 to 6 weeks intervals until either CR or progressive disease could be determined. Posttreatment surveillance after CR consisted of clinical review and 6-monthly imaging.

The primary endpoints were CR and safety. CR was achieved if, at any time point, there was no residual disease clinically or on FDG PET, or residual imaging findings which were biopsy negative, or if repeat PET imaging demonstrated resolution or stability of residual abnormalities. Adverse events (AEs) were graded according to National Cancer Institute Common Toxicity Criteria for AEs version 4.03.

Key secondary endpoints were overall survival (OS), progression-free survival (PFS), and time to progression.

Statistical analysis

Based on our previous prospective study showing a CR rate of 50% with chemoradiation,¹⁴ the null hypothesis was a response rate of 50% or less. With the addition of durvalumab, we hypothesized that the response rate would be 70% or more, with significance tested using a 2-sided binomial test with a target significance level of 0.10 and 80% power. A 2-stage minimax design was applied, and we calculated that an initial sample size of 15 patients would be required for the first stage. If at least 8 of the first 15 patients responded, an additional 13 patients would be required for the second stage, subject to feasibility. Of the total of 28 patients, 18 or more must respond for the treatment to be considered better than the current treatment.

The primary analysis was assessed in the intention-to-treat population. The primary outcome of CR was expressed as a percentage with 95% CIs compared with the hypothesized null effect of 0.5 using an exact binomial test. Time-to-event outcomes were defined from the time of enrolment and estimated with the Kaplan-Meier method. Analyses were conducted using Stata (version 18).¹⁷

Results

Patients

From May 2019 through June 2021, we enrolled 15 patients with unresectable cSCC. The median age was 75. Nodal disease was present in 14 (93%) patients. One patient had primary disease arising from the lower limb, 1 from the chest wall, and the remainder from the head and neck region. Disease was deemed unresectable due to a low likelihood of achieving macroscopic clearance in 12 (80%) and medical unsuitability in 3 (20%). Baseline demographics are presented in Table 1.

Table 1.

Patient characteristics

Characteristic	Patients (n = 15)
Median age—years (range)	75 (58-85)
Sex—no. (%)  Female  Male	3 (20) 12 (80)
ECOG performance status—no. (%)  0  1	7 (47) 8 (53)
Primary site—no. (%)  Head and neck  Other	13 (87) 2 (13)
Treatment site—no. (%)  Local only  Nodal only  Local and nodal  Perineural	1 (7) 8 (53) 6 (40) 0 (0)
T stage (AJCC 7th edition)—no. (%)  0  1  2  3  4  Unknown	2 (13) 1 (7) 1 (7) 3 (20) 1 (7) 7 (47)
N stage (AJCC 7th edition)—no. (%)  0  1  2  3	1 (7) 0 (0) 10 (67) 4 (27)
Stage (AJCC 7th edition)—no. (%)  I  II  III  IV	0 (0) 1 (7) 0 (0) 14 (93)
Chemotherapy  Cisplatin  Carboplatin	3 (20) 12 (80)

Abbreviation: AJCC = American Joint Committee on Cancer; ECOG = Eastern Cooperative Oncology Group.

Response

All 15 patients were assessable for response. CR was observed in 9 of 15 (60%; 95% CI, 32%-84%) patients. Two of these were achieved at the 19-week follow-up, an additional 2 at the 23-week follow-up, and the remaining 5 between 31 and 41 weeks from initial treatment. At data cut off, all patients who achieved a CR remain alive and disease free apart from 1 patient who died due to pancreatic cancer, with a median follow-up of 39 months (26-44).

The 6 patients who did not achieve a CR were all eventually deemed to have progressive disease; 2 had progressive disease at the 19-week follow-up, and the longest time from initial treatment to determining progressive disease was 68 weeks. For the patients with progressive disease, 5 died of SCC and 1 received combination immunotherapy on another clinical trial.

Administration and safety

All 15 patients completed the prescribed course of radiation therapy. The recommended 70 Gy in 35 fractions was delivered in 13 patients; 2 had planned variations. One patient had bulky unresectable nodal disease in the neck and subsequently had dermatitis, weight loss, and deconditioning from treatment; thus, radiation therapy was ceased at 68 Gy in 34 fractions. The other patient had primary cSCC at the right lower leg (10 cm primary) with nodal metastases to the ipsilateral popliteal, inguinal, pelvic, and small-volume, low paraortic lymph nodes. She was prescribed and delivered 66 Gy in 33 fractions to the right lower leg due to the lower limb location and large treatment volume.

Eight patients completed all planned doses of chemotherapy. Five patients had treatment withheld for 2 to 4 weeks because of thrombocytopenia or neutropenia. One patient was admitted because of confusion and deemed too unwell to receive chemotherapy. One patient had systemic therapy ceased after 4 weeks because of grade 3 pneumonitis attributed to durvalumab.

At least 5 cycles of durvalumab, to reach the first 19-week assessment, were delivered to 12 patients. There was one treatment-related serious AE of grade 3 pneumonitis, onset after 4 weeks, attributed to durvalumab. This patient completed radiation therapy but did not achieve a CR and died due to SCC at 6 months from initial treatment. Two patients voluntarily withdrew from treatment; 1 received 2 cycles with a partial but incomplete response after chemoradiation and died 22 months from initial treatment, and 1 received 3 cycles with a CR. The median number of durvalumab cycles was 7, with the highest number of cycles being 10 before progressive disease was determined. Table 2 outlines the treatment received by each patient.

Table 2.

Treatment summary

				Chemotherapy
	Sites treated	Radiation dose (Gy)—planned	Radiation dose (Gy)—received	Agent	No. cycles	Durvalumab—no. cycles	CR
1	Nodal only	70	68	Cisplatin	4	1	No
2	Local and nodal	70	70	Carboplatin	7	7	No
3	Local and nodal	70	70	Carboplatin	6	2	No
4	Nodal only	70	70	Cisplatin	4	8	Yes
5	Local and nodal	66	66	Carboplatin	7	5	No
6	Nodal only	70	70	Carboplatin	2	10	No
7	Nodal only	70	70	Carboplatin	5	7	Yes
8	Nodal only	70	70	Carboplatin	7	3	Yes
9	Nodal only	70	70	Carboplatin	7	9	Yes
10	Local and nodal	70	70	Carboplatin	4	5	Yes
11	Local only	70	70	Carboplatin	7	9	Yes
12	Local and nodal	70	70	Carboplatin	3	9	Yes
13	Local and nodal	70	70	Carboplatin	7	7	Yes
14	Nodal only	70	70	Cisplatin	7	7	No
15	Nodal only	70	70	Carboplatin	4	9	Yes

Abbreviation: CR = complete response.

All patients experienced at least 1 AE during treatment. Apart from pneumonitis, there was an additional serious AE of pulmonary embolism on completion of chemoradiation. Chemotherapy and radiation therapy side effects were otherwise as expected and summarized in Table 3. There were no deaths due to treatment-related AEs.

Table 3.

Adverse effects

	Number of patients (%)
Adverse event	Any grade	Immune related	Grade ≥ 3
Any	15(100)		9 (60)
Serious adverse event	9 (60)		3 (20)
Immunotherapy-related event	11 (73)		1 (7)
Dermatitis	14 (93)	2 (13)	2 (13)
Fatigue	11 (73)	6 (40)	0
Constipation	8 (53)	0	0
Nausea	7 (47)	1 (7)	0
Thrombocytopenia	7 (47)	0	0
Anemia	7 (47)	0	1 (7)
Xerostomia	5 (33)	0	0
Dysgeusia	4 (27)	0	0
Rash	3 (20)	2 (13)	0
Tinnitus	3 (20)	0	0
Dysphagia	3 (20)	0	1 (7)
Neutropenia	2 (13)	0	2 (13)
Diarrhea	2 (13)	0	0
Neuropathy	2 (13)	0	0
Mucositis	2 (13)	0	0
Pneumonitis	1 (7)	1 (7)	1 (7)
Pulmonary embolus	1 (7)	0	1 (7)
Increased ALT/AST	1 (7)	0	0
Increased creatinine	1 (7)	0	0
Hearing impairment	1 (7)	0	0
Confusion	1 (7)	0	1 (7)

Abbreviation: ALT = Alanine Aminotransferase; AST = Aspartate Aminotransferase.

Survival

Median follow-up time was 33 months (IQR, 20-41; range, 6-44). Median overall and PFS were not reached. Estimated OS at 12, 24, and 36 months were 86% (95% CI, 55%-96%), 65% (95% CI, 35%-84%), and 57% (95% CI, 27%-78%) as shown in Fig. 1. Estimated PFS rates at 12, 24, and 36 months were 86% (95% CI, 55%-96%), 65% (95% CI, 35%-84%), and 65% (95% CI, 35%-84%). For the 6 patients who had progressive disease or died from cSCC, the median time to progression was 8 months (range, 5-23).

Figure 1.

Overall survival (A) and progression-free survival (B).

Discussion

The CRIO study is the only prospective study of the combined approach of chemoradiation and immunotherapy for unresectable cSCC. A CR was achieved in 60% of patients and none of these patients had experienced recurrent disease after a median follow-up of 39 months. All patients who did not achieve a CR had progressive disease; 1 went on to receive combination immunotherapy on a clinical trial and remains alive and disease free after more than 3 years, and the other 5 died due to disease progression.

This result builds on our prospective study on concurrent chemoradiation, which reported a CR rate of 53%, 3-year OS of 55%, and disease-free survival of 30% with 66% of patients having unresectable stage IV disease.¹⁴ A recent retrospective study on definitive radiation therapy alone reported a median OS of 21 months and durable disease control in just over half of patients; however, only 17% of patients had stage IV disease and the definition of local control was not clearly defined.⁶ Another found an OS of 19 months and 3-year disease-specific survival of 38%, but excluded patients with nodal metastases.⁷

The safety and tolerability of this treatment regimen were as expected for concurrent chemoradiation¹⁴ and single-agent checkpoint inhibitor immunotherapy. There was 1 case of grade ≥3 toxicity related to immunotherapy as grade 3 pneumonitis, which resolved with treatment, but the patient died due to disease progression after 6 months. Two patients withdrew from completing 5 cycles of durvalumab before the 19-week assessment for reasons unrelated to toxicity. Most patients (80%) received carboplatin, with only 8 (53%) completing all planned cycles of chemotherapy, reflective of the elderly, comorbid patient population seen in locoregionally advanced cSCC.

Despite being a tertiary surgical referral center, enrolment for the first part of this study required 2 years owing to the limited population for which nonsurgical management was recommended. At the time of study design, checkpoint inhibitor monotherapy for unresectable or metastatic cSCC was not yet considered standard of care. Two PD-1 inhibitors, cemiplimab and pembrolizumab, have since been shown to achieve high radiological response rates of up to 60%, many of which are durable.11, 12, 13 Thus, immunotherapy alone is often used in the setting of unresectable disease, although radiation therapy might still be a viable option for some patients.

There has also been interest in neoadjuvant checkpoint inhibitor therapy for resectable cSCC. There have been single-arm phase 2 trials of preoperative treatment for high-risk but resectable cSCC with both cemiplimab and pembrolizumab showing pathological CR rates of over 50%.^18,19 The phase 2 trial DESQUAMATE demonstrated the potential to avoid surgery and radiation therapy altogether for patients who achieve an excellent response with neoadjuvant pembrolizumab.²⁰ Additionally, 2 phase 3 studies evaluating checkpoint inhibitor immunotherapy following radiation therapy in the high-risk adjuvant setting were recently published. The C-POST trial showed a significantly longer disease-free survival with the addition of cemiplimab to surgery and radiation therapy,²¹ although no significant benefit was demonstrated with pembrolizumab in KEYNOTE-630.²²

The ongoing phase 3 NRG HN014 study is investigating a response-adapted approach to high-risk cSCC with preoperative cemiplimab, with the omission of postoperative radiation therapy and immunotherapy for patients who achieve a pathological CR.²³

This paradigm shift to employing PD-1/PD-L1 inhibitors as either palliative monotherapy or perioperative treatment for advanced cSCC led us not to proceed to the second stage of this trial as initially planned, despite not accepting the null hypothesis at the end of stage one. We recognize that this novel combination of chemoradiation with immunotherapy was conceptually stimulating at the time while the protocol was developed. We are not exploring the CRIO approach further unless longer-term follow-up for immunotherapy reports unexpected late recurrence or progression.

One limitation was the definition of unresectable disease, which was not strictly defined but deemed as such by an experienced multidisciplinary clinic consisting of multiple surgical specialties, subspecialist radiation, and medical oncologists, and supporting nursing and allied health staff.

Another major limitation was the definition of CR. Both radiation therapy and immunotherapy pose challenges with radiological assessment of disease response after treatment. We acknowledge the partly subjective nature of our definition of CR, as well as the role of FDG PET in assessing cSCC although there has been retrospective data to support an increased accuracy of disease response assessment with FDG PET over conventional CT.^24,25 Our previous experience with the study found residual radiological disease was common after treatment, which influenced the definition of CR in this study. We also acknowledge the ambiguity associated with an undefined timepoint at which to declare response. Patients eventually had to be assigned to having a CR or progressive disease. Of the 9 patients who achieved a CR, only 2 reached this at the first assessment at 19 weeks, with the rest taking up to 41 weeks. The longest follow-up to determine progressive disease was 68 weeks. This aimed at pragmatically estimating the proportion of patients who derived durable disease control with the study intervention, and the lack of disease relapse with a median follow-up of 39 months in all patients ultimately assessed to have had a CR lends credibility to our approach.

Conclusion

This small, single-institution study provides prospective data for a unique approach of chemoradiation in combination with checkpoint inhibitor therapy for patients with unresectable cSCC, a patient group for whom there is little information to guide management. This regimen appears to be associated with a high CR rate, responses were durable, and treatment was feasible with acceptable toxicity.

Disclosures

Charles Lin reports that financial support was provided by AstraZeneca. Brett G.M. Hughes reports a relationship with AstraZeneca, Merck Sharpe & Dohme, Bristol-Myers Squibb, Pfizer, Roche, Regeneron, Sanofi, Amgen, and Eisai that includes: consulting or advisory. Gary Ng reports a relationship with Pfizer and Daiichi Sankyo that includes: speaking and lecture fees. Gary Ng reports a relationship with AstraZeneca that includes: speaking and lecture fees and travel reimbursement. Gary Ng reports a relationship with Merck Sharpe & Dohme, Merck, and Roche that includes: travel reimbursement.. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.