Introduction

Since the introduction of immune checkpoint inhibitors (ICIs) as standard of care for patients with metastatic melanoma (MM), survival rates have improved significantly in pivotal phase III trials and in real-life patients.1 2 For patients with tumors harboring an activating BRAF mutation, BRAF/MEK inhibitors have also been shown to increase overall survival (OS).3

Nevertheless, a substantial proportion of patients either do not benefit from treatment or eventually relapse, highlighting the need for additional treatment options. Recently, adoptive cell therapy (ACT) using tumor-infiltrating lymphocytes (TILs) showed superior progression-free survival (PFS) compared with ipilimumab monotherapy in patients with MM in an academically driven randomized phase III trial.1 Further, a commercially available TIL product achieved United States Food and Drug Administration (FDA) approval in February 2024 in patients with MM previously treated with an anti-PD-1 antibody, and in the presence of a BRAF mutation, a BRAF inhibitor with or without a MEK inhibitor.2 The FDA approval was granted based on phase II data.3

Previous reports have shown that PFS and OS after ACT using TILs correlate well with objective response according to Response Evaluation Criteria In Solid Tumors (RECIST). Especially complete responses (CR) are durable and associated with long-term survival.4 However, CR sometimes takes long time to develop and a substantial subgroup of patients achieving an initial partial response (PR) will eventually progress. Thus, exploring ways to predict prognosis early after treatment is highly warranted as identifying patients at high risk of imminent progression would allow for closer follow-up and early onset of additional treatment.

Metabolic response measured by [18F]2-fluoro-2-deoxy-d-glucose positron emission tomography (FDG-PET) scans has been shown to be a valuable parameter to predict clinical outcomes of patients with MM treated with ICI.5 6

In this study, we investigated whether early metabolic response provides prognostic information in patients with MM treated with ACT using TILs.

MethodsStudy design and patients

The study population consists of patients with MM treated with TIL therapy in clinical trials spanning from July 2011 to June 2022 (NCT00937625, NCT02379195, NCT02354690 and NCT02278887) at the National Center for Cancer Immunotherapy, Department of Oncology, Copenhagen University Hospital, Herlev.

All patients were treated with 7 days of lymphodepleting chemotherapy with cyclophosphamide and fludarabine, infusion of autologous TILs and interleukin-2 (IL-2). In three trials, patients were treated with low-dose IL-2 subcutaneously or with a continuous intravenous IL-2 dosing regimen (decrescendo regimen), as previously described.7–10 In one trial, patients were treated with high-dose bolus IL-2 intravenously every 8 hours, as previously described.11

FDG-PET scans were performed with diagnostic CT scans before therapy, 6–8 weeks post-TIL infusion, and 12–16 weeks post-TIL infusion. This analysis included only patients with an available baseline scan and at least one evaluation FDG-PET/CT follow-up scan 6–8 weeks and/or 12–16 weeks post-TIL infusion. FDG-PET/CT scans were performed as a standard of care.

Radiologic responses were assessed according to RECIST V.1.112 and metabolic responses according to Positron Emission Tomography Response Evaluation Criteria In Solid Tumors (PERCIST).13 An experienced and blinded nuclear physicist reviewed metabolic response for consistency. Unless otherwise specified, metabolic response is reported as overall metabolic response at the time of second post-treatment scan.

Data cut-off for this analysis was December 31, 2023.

Statistical analysis

OS and PFS curves were computed according to the Kaplan-Meier method using GraphPad Prism V.10.1.2 software. P values of comparisons between survival curves were performed using the log-rank Mantel-Cox test. A p<0.05 was considered significant.

Concordance between PERCIST and RECIST response categories was calculated using Cicchetti-Allison-weighted kappa statistics. A kappa of 0.81–1.00 was considered as almost perfect agreement, 0.61–0.80 as substantial agreement, 0.41–0.60 as moderate agreement, 0.21–0.40 as fair agreement, 0–0.20 as slight agreement, and <0.00 as poor agreement.14

ResultsPatient characteristics

Baseline and at least one evaluation FDG-PET/CT scan were available for 60 patients and the median follow-up time was 98.7 months, with a minimum follow-up of 18 months. The majority of the patients excluded from this analysis were in the early trials due to not having FDG-PET/CT scans available. See online supplemental figure 1 for consort diagram and overview over which IL-2 regimen was administered. The treatment landscape changed during the period of this study with most of the early patients having received prior ipilimumab (38/60 patients) and most of the later patients having received prior anti-PD-1 blockade (35/60 patients). In total, 93% of the patients had any type of ICI in either adjuvant or metastatic setting (see online supplemental table 2).

RECIST and PERCIST responses after TIL therapy

Within the first two evaluation scans after TIL therapy, only two patients achieved a CR, 22 patients had a PR, 12 patients had stable disease and 24 had progressive disease (PD). RECIST responses correlate well with OS, thus, obtaining an early objective response (CR or PR) is associated with significantly longer OS (mOS not reached, range 12.8–149.8+months) compared with not obtaining an objective response (mOS 13.1 months, range 3.2–143.6+; p<0.0001) (figure 1A and C and online supplemental figure 2A).

Figure 1

(A–D) Overall survival (upper panels) and progression-free survival (lower panels) stratified according to Response Evaluation Criteria In Solid Tumors (RECIST) (left panels) and Positron Emission Tomography Response Evaluation Criteria In Solid Tumors (PERCIST) (right panels). CR, complete response; CMR, complete metabolic response; PD, progressive disease; PMD, progressive metabolic disease; PR, partial response; PMR, partial metabolic response; SD, stable disease; SMD, stable metabolic disease.

In addition, PERCIST response was evaluated (figure 1B,D); 12 patients achieved complete metabolic response (CMR) with the first two evaluation scans, 14 had partial metabolic response (PMR), 6 had stable metabolic disease (SMD) and 28 had progressive metabolic disease (PMD). Obtaining a CMR is associated with significantly longer OS (mOS not reached, range 18.3+−149.8+) compared with not obtaining a CMR (non-CMR) (mOS 16.1 months, range 3.2–143.6+; p<0.0001) (online supplemental figure 2B). The PFS curves subcategorized according to RECIST and PERCIST visually differ due to the separation of patients into aforementioned subgroups (figure 1C,D) especially due to more patients achieving CMR than CR.

Predictive value of early metabolic response after TIL therapy

10 of the 12 patients who achieved and maintained CMR within the second evaluation scan later developed CR and 2 had durable deep PRs according to RECIST. One patient died of another cause and melanoma-specific survival after CMR at second evaluation is 100%.

In the PR group, 10 patients concurrently achieved CMR, and 12 did not. This translated into a statistically significant difference in mOS for the PR/non-CMR group of 28.7 months vs not reached (range 32.2–145.4+) for the PR/CMR group (figure 2A; p=0.0016). For the PR/non-CMR group, mPFS was 7.8 months vs not reached (range 20.7–123.4+) for the PR/CMR group (figure 2B; p<0.0001).

Figure 2

(A, B) Showing overall survival (OS) (left panel) and progression-free survival (PFS) (right panel) for patients with partial response (PR) stratified by obtaining early complete metabolic response (CMR) or not (non-CMR). mOS, median OS.

For the entire cohort, not obtaining a CMR was associated with a short mPFS of only 3.2 months and 97.9% (47/48) of patients progressed within 12 months.

Predictive value of a scan 6–8 weeks post-treatment scan

In this study, early treatment evaluation was performed at 6–8 weeks and 12–16 weeks post-treatment. It was addressed whether the early scan after 6–8 weeks added clinical useful information regarding durability of treatment efficacy.

At the time of the first scan, 11 patients had SMD and 17 had PMD, which was associated with a very poor mPFS of 2.7 months (range 1.0–5.8) and mOS of 13.1 months (range 3.2–143.6+). None of these 28 patients obtained a better response at a later time point, that is, pseudoprogression was not observed. In comparison, only nine patients had PD according to RECIST at the first scan.

As expected, PMR and CMR were associated with a better prognosis, however, the difference between the two groups was less pronounced compared with the time of the second scan (data not shown).

Concordance of RECIST and PERCIST response

Overall agreement of paired response categories (CR and CMR, PR and PMR etcetera) was moderate (61.7%) with the lowest agreement in the CR/CMR category at 28.6% and the highest in the PD/PMD category at 79.3% (online supplemental table 1 and figure 3).

Figure 3

Graphical presentation of mismatch between Response Evaluation Criteria In Solid Tumors (RECIST) and Positron Emission Tomography Response Evaluation Criteria In Solid Tumors (PERCIST) response classification. CR, complete response; CMR, complete metabolic response; PD, progressive disease; PMD, progressive metabolic disease; PR, partial response; PMR, partial metabolic response; SD, stable disease; SMD, stable metabolic disease.

In short, adding FDG-PET to conventional CT allows for faster and better separation of patients with good prognosis and long-term treatment response (ie, CMR) from patients with poor outcomes (ie, SMD and PMD). More patients are classified in a poorer response category when using metabolic responses. In total, progression was indicated by metabolic progression prior to structural progression in 25% of patients (15/60) either at first or second post-treatment scan. In only three patients, structural progression was observed before metabolic progression, and all three patients had SMD.

Metabolic progression was in a few cases used to guide resection of oligometastatic disease. Thus, one patient had suspected relapse of a single lymph node with increased metabolism which was not structurally enlarged on the CT scan. The lesion was surgically removed at 13 months post-treatment, and the presence of viable malignant cells was confirmed by histology. This patient is still without evidence of disease for more than 11 years without further treatment. Another patient had a single lesion removed due to PMD on two consecutive scans while retaining PR structurally. The presence of viable malignant cells was confirmed by histology. This patient is also alive and without evidence of disease for more than 11 years.

Discussion

In patients treated with ACT using TILs evaluation of early metabolic response allows for better prediction of durability of response and thus prognostication, with CMR being a strong predictor of long-term PFS and OS. This is in line with previous reports on metabolic response after immunotherapy using ICI.5 6

While adding a FDG-PET scan to the diagnostic CT scan does increase time and costs of scans and also the radiation dose of each scan, these considerations are easily justified by the opportunity to tailor patient-specific follow-up through better stratification of prognosis after treatment. Patients achieving CMR could potentially undergo less extensive follow-up, that is, longer intervals between scans or shorter follow-up, or may not need additional scans at all if also achieving CR, given the rarity of relapse after CR.4 15 Conversely, even in the absence of structural progression, patients with SMD or PMD at the very early scan 6–8 weeks after treatment or non-CMR status at 12–16 weeks after treatment are at high risk of progression and should be closely monitored.

At our site, most advanced melanoma patients are followed using FDG-PET/CT scans as a standard of care. Thus, at our site, there was no extra cost performing this study, however, this might differ between sites and countries.

The radiation dose of a diagnostic CT of thorax and abdomen is approximately 14 mSv. Adding FDG-PET increases the radiation dose to approximately 17 mSv. After 2 years of follow-up and a total of 10 scans, the additional radiation dose using FDG-PET/CT scans is equivalent of two extra CT scans. However, FDG-PET scans can be performed with a low-dose CT, lowering the radiation dose to approximately 12 mSv. According to our national guidelines, low-dose FDG-PET/CT scan is used for follow-up if the patient has CR for more than 1 year. For instance, if a patient obtains a CR after 6 months and switches to low-dose FDG-PET/CT after another year of follow-up, then the total radiation dose of the two modalities will be equivalent after 5 years of follow-up (a total of 18 scans).

The high-risk status of the patients with suboptimal metabolic response also opens a window of opportunity to add-in rescue treatment. In a few patients included in this analysis, surgical removal of oligometastatic lesions after metabolic, but not structural, progression has resulted in patients living for more than 10 years with no further evidence of disease progression or relapse. Although this is case based, it suggests a potential survival benefit of surgery in selected patients, in line with conclusions of Klemen et al.16 Further, additional consolidation therapies, for example, personalized vaccines or checkpoint inhibitors, could be considered. In line with this, ongoing trials are presently testing whether addition of a dendritic cell vaccine improves the efficacy of TIL treatment (ClinicalTrialgs.gov identifiers NCT01946373, NCT00338377), or if insufficient response after TIL treatment can be rescued with the addition of anti-PD-1 blockade (NCT03475134), even in anti-PD-1 refractory patients.

While the TILs used in the clinical trials included in this analysis are grown in a similar way to the TILs in the FDA-approved product lifileucel (eg, extraction of TILs during in vitro culture without otherwise comparing culture methods), they are not identical. The lifileucel infusion product is cryopreserved and shipped from a centralized production facility, while our TILs could be cryopreserved during production before the Rapid Expansion Process (35% of products, data not shown), but the infusion product was always fresh. Furthermore, patients in our study were treated with various IL-2 regimens. Thus, the findings of this study need to be validated in cohorts of ICI-refractory patients treated according to the currently approved regimen.

In conclusion, early CMR is a strong predictor of durability of response and long-term survival after TIL therapy in MM patients. Furthermore, FDG-PET/CT scans appear superior to CT scans for providing early prognostic information after TIL therapy in melanoma and can be used to tailor patient-specific management and follow-up strategies.

Data availability statement

Data are available on reasonable request. Data are available on reasonable request from the corresponding author.

Ethics statementsPatient consent for publication

Not applicable.

Ethics approval

This study involves human participants and was approved by the Ethical Committee of the Capital region of Denmark reference number Journal-nr.: H-20070020. Participants gave informed consent to participate in the study before taking part.