Early Recurrence in Completely Resected IIIB and IIIC Melanoma Warrants Restaging Prior to Adjuvant Therapy

After signing informed consent, patients were screened for eligibility in a placebo-controlled randomized trial (NCT02993315) investigating adjuvant dendritic cell vaccination. The protocol has been approved by the national review committee (Central Committee on Research Involving Human Subjects) and is in concordance with the Declaration of Helsinki and Good Clinical Practice. Eligible patients were adults with stage IIIB or IIIC [American Joint Committee on Cancer (AJCC) 7th edition]2 cutaneous melanoma within 12 weeks after complete radical lymph node dissection (RLND) and after recovery from the surgery. The protocol was amended after publication of the MSLT-II trial results, which showed no survival benefit of completion lymph node dissection after removal of microscopic metastasis with sentinel node biopsy (SNB) when compared with nodal surveillance.26 After amendment, patients with microscopic disease could be included after SNB and additional completion lymph node dissection was no longer required. Macrometastasis was defined as a palpable node or as a nonpalpable node of at least 15 mm in short axis on CT, a PET-positive node, or one or more foci of melanoma of at least 1 cm in diameter in the pathology report. Patients with completely resected in-transit and/or satellite metastasis, an unknown primary tumor, and (planned) adjuvant radiotherapy could be included. In addition, absence of distant metastasis had to be documented by ceCT of the chest, abdomen, and pelvis or whole-body ¹⁸F-FDG PET scan combined with CT (¹⁸F-FDG PET/CT) within 6 weeks before inclusion in our trial. In patients with head or neck melanoma, additional ceCT of the neck was obligatory. Imaging of the brain was performed in case of clinical suspicion of brain metastasis. Exclusion criteria included autoimmune disease (except for skin disease, hypothyroidism after autoimmune thyroiditis, and type 1 diabetes mellitus), a second malignancy in the last 5 years (except for adequately treated carcinoma in situ and basal or squamous cell carcinoma of the skin), concomitant use of oral or intravenous immunosuppressive drugs, and uncontrolled infectious disease.

Between November 2016 and July 2018, 120 patients were screened for eligibility. Baseline characteristics are presented in Table 1. Median age was 54 (range 27–79) years, and 76 (63%) of patients were male. Sixty-nine (58%) and 51 (43%) patients were diagnosed with stage IIIB and IIIC melanoma, respectively. Twenty-one (18%) patients had completely resected in-transit metastasis, and nine (8%) patients presented with nodal metastasis from an unknown primary tumor. Baseline characteristics of patients with and without recurrent disease during screening are presented in Table 1. No statistically significant differences between groups were present.

Melanoma metastasis was detected in 22 (18%) of 120 patients (Fig. 1), corresponding to a number needed to screen of 5.45 to detect one patient with recurrent disease. Thirteen (59%) patients were identified with distant metastasis, while in the remaining nine (41%) patients, metastasis was locoregionally located.

Five (23%) recurrences were found based on symptoms or physical examination (symptomatic recurrence); in three patients, in-transit metastasis was noticed by the patient (n = 1) or physician (n = 2), and another patient discovered a local recurrence at the site of the resected primary melanoma. Of these four patients with symptomatic locoregional relapse, two showed detectable distant metastatic disease on ceCT scan. The fifth patient developed back pain, which was suspicious for bone metastasis and confirmed by ceCT imaging. Seventeen (77%) relapses were asymptomatic and initially detected by imaging, corresponding to a number of asymptomatic patients needed to screen of 6.76. One of these patients presented with atypical, very small pulmonary nodules before RLND. Another patient showed atypical/nonspecific hypodense liver lesions of maximum 10 mm on preoperative ceCT scan, and these lesions were identified as liver metastases during screening ceCT after a 12-week interval. Serum LDH level was not a sensitive parameter for recurrent disease, since only 4 (18%) out of 22 relapsed patients had elevated LDH. All four patients had distant metastasis, and two of them were asymptomatic.

Before referral to our trial, metastasis had been excluded with ¹⁸F-FDG PET/CT (85%) or ceCT (15%) in 115 patients. Of the five patients in whom metastasis had not been excluded prior to screening, four had resected micrometastasis in the pathology report and one patient had macrometastatic disease. However, in all patients presenting recurrent disease during screening, distant metastasis had been excluded on imaging prior to start of screening for eligibility (Fig. 2). For this group with early relapse, prior imaging was done using ¹⁸F-FDG PET/CT in 20 patients (91%) and ceCT in the remaining 2 patients.

To screen for eligibility, 110 (96%) patients had standard ceCT. In the remaining five (4%) patients, imaging was performed by ¹⁸F-FDG PET/CT. Relapse during screening was detected by ceCT in all cases. In five patients, imaging was not repeated during screening, since the start of experimental adjuvant therapy was within 6 weeks after prior imaging excluding distant metastasis.

The median interval between imaging during screening and previous imaging was 10.2 (range 5.7–20.9) weeks in recurrent patients. The median interval between complete resection and detection of recurrent disease was 7.4 (range 4.3–10.7) weeks. In patients without recurrent disease, these intervals were not significantly different, with a median interval between scans of 11.1 (range 5.6–27.7) weeks and an interval between resection and imaging of 7.3 (range 4.0–11.0) weeks. Figure 3 shows examples of patients with asymptomatic recurrent disease.

Nine patients showed locoregional metastasis, of whom eight were referred for surgical resection with curative intent. One patient had no evidence of disease after adjuvant radiotherapy, therefore planned surgery was cancelled. This patient was disease free during 13 months of follow-up, then relapsed. Of the seven reoperated patients, six developed recurrent disease. In two of them, distant metastasis occurred within 1 month after resection of the recurrent local disease. In four patients, the interval from resection to recurrent disease was 6, 6, 8, and 9 months. The last reoperated patient is still recurrence free after 10 months of follow-up. In the remaining patient, locoregional recurrence consisted of irresectable in-transit metastasis, for which treatment with anti-PD-1 antibodies was initiated.

Of the 13 patients with distant metastasis, first-line treatment consisted of anti-PD-1 antibodies in three patients, three patients started with combined immune checkpoint inhibition, and in six patients treatment with targeted therapy was initiated. One patient underwent metastasectomy of a solitary liver metastasis.