Explainable artificial intelligence in skin cancer recognition: A systematic review

BACKGROUND

Due to their ability to solve complex problems, deep neural networks (DNNs) are becoming increasingly popular in medical applications. However, decision-making by such algorithms is essentially a black-box process that renders it difficult for physicians to judge whether the decisions are reliable. The use of explainable artificial intelligence (XAI) is often suggested as a solution to this problem. We investigate how XAI is used for skin cancer detection: how is it used during the development of new DNNs? What kinds of visualisations are commonly used? Are there systematic evaluations of XAI with dermatologists or dermatopathologists?

METHODS

Google Scholar, PubMed, IEEE Explore, Science Direct and Scopus were searched for peer-reviewed studies published between January 2017 and October 2021 applying XAI to dermatological images: the search terms histopathological image, whole-slide image, clinical image, dermoscopic image, skin, dermatology, explainable, interpretable and XAI were used in various combinations. Only studies concerned with skin cancer were included.

RESULTS

37 publications fulfilled our inclusion criteria. Most studies (19/37) simply applied existing XAI methods to their classifier to interpret its decision-making. Some studies (4/37) proposed new XAI methods or improved upon existing techniques. 14/37 studies addressed specific questions such as bias detection and impact of XAI on man-machine-interactions. However, only three of them evaluated the performance and confidence of humans using CAD systems with XAI.

CONCLUSION

XAI is commonly applied during the development of DNNs for skin cancer detection. However, a systematic and rigorous evaluation of its usefulness in this scenario is lacking.

Clinical determinants of long-term survival in metastatic uveal melanoma

This study aimed to identify prognostic factors in patients with metastatic uveal melanoma (UM) that were associated with long-term survival in a real-world setting. A total of 94 patients with metastatic UM were included from German skin cancer centers and the German national skin cancer registry (ADOReg). Data were analyzed for the response to treatment, progression-free survival, and overall survival (OS). Prognostic factors were explored with univariate Cox regression, log-rank, and χ2-tests. Identified factors were subsequently validated after the population was divided into two cohorts of short-term survival (< 2 years OS, cohort A, n = 50) and long-term survival (> 2 years OS, cohort B, n = 44). A poor ECOG performance status (hazard ratio [HR] 2.0, 95% confidence interval [CI] 1.0–3.9) and elevated serum LDH (HR 2.0, 95% CI 1.0–3.8) were associated with a poor OS, whereas a good response to immune checkpoint blockade (ICB, p < 0.001), radiation therapy (p < 0.001), or liver-directed treatments (p = 0.01) were associated with a prolonged OS. Long-term survivors (cohort B) showed a higher median number of organs affected by metastasis (p < 0.001), while patients with liver metastases only were more common in cohort A (40% vs. 9%; p = 0.002). A partial response to ICB was observed in 16% (12/73), being 21% (8/38) for combined ICB, 17% (1/6) for single CTLA4 inhibition, and 10% (3/29) for single PD1 inhibition. One complete response occurred in cohort B with combined ICB. We conclude that the response to ICB and the presence of extrahepatic disease were favorable prognostic factors for long-term survival.

Overall Survival Benefit with Tebentafusp in Metastatic Uveal Melanoma

BACKGROUND

Uveal melanoma is a disease that is distinct from cutaneous melanoma, with a low tumor mutational burden and a 1-year overall survival of approximately 50% in patients with metastatic uveal melanoma. Data showing a proven overall survival benefit with a systemic treatment are lacking. Tebentafusp is a bispecific protein consisting of an affinity-enhanced T-cell receptor fused to an anti-CD3 effector that can redirect T cells to target glycoprotein 100-positive cells.

METHODS

In this open-label, phase 3 trial, we randomly assigned previously untreated HLA-A*02:01-positive patients with metastatic uveal melanoma in a 2:1 ratio to receive tebentafusp (tebentafusp group) or the investigator's choice of therapy with single-agent pembrolizumab, ipilimumab, or dacarbazine (control group), stratified according to the lactate dehydrogenase level. The primary end point was overall survival.

RESULTS

A total of 378 patients were randomly assigned to either the tebentafusp group (252 patients) or the control group (126 patients). Overall survival at 1 year was 73% in the tebentafusp group and 59% in the control group (hazard ratio for death, 0.51; 95% confidence interval [CI], 0.37 to 0.71; P<0.001) in the intention-to-treat population. Progression-free survival was also significantly higher in the tebentafusp group than in the control group (31% vs. 19% at 6 months; hazard ratio for disease progression or death, 0.73; 95% CI, 0.58 to 0.94; P = 0.01). The most common treatment-related adverse events in the tebentafusp group were cytokine-mediated events (due to T-cell activation) and skin-related events (due to glycoprotein 100-positive melanocytes), including rash (83%), pyrexia (76%), and pruritus (69%). These adverse events decreased in incidence and severity after the first three or four doses and infrequently led to discontinuation of the trial treatment (2%). No treatment-related deaths were reported.

CONCLUSIONS

Treatment with tebentafusp resulted in longer overall survival than the control therapy among previously untreated patients with metastatic uveal melanoma. (Funded by Immunocore; ClinicalTrials.gov number, NCT03070392; EudraCT number, 2015-003153-18.).

Immune Checkpoint Blockade for Metastatic Uveal Melanoma: Patterns of Response and Survival According to the Presence of Hepatic and Extrahepatic Metastasis

BACKGROUND

Since there is no standardized and effective treatment for advanced uveal melanoma (UM), the prognosis is dismal once metastases develop. Due to the availability of immune checkpoint blockade (ICB) in the real-world setting, the prognosis of metastatic UM has improved. However, it is unclear how the presence of hepatic and extrahepatic metastasis impacts the response and survival after ICB.

METHODS

A total of 178 patients with metastatic UM treated with ICB were included in this analysis. Patients were recruited from German skin cancer centers and the German national skin cancer registry (ADOReg). To investigate the impact of hepatic metastasis, two cohorts were compared: patients with liver metastasis only (cohort A, n = 55) versus those with both liver and extra-hepatic metastasis (cohort B, n = 123). Data were analyzed in both cohorts for response to treatment, progression-free survival (PFS), and overall survival (OS). The survival and progression probabilities were calculated with the Kaplan-Meier method. Log-rank tests, χ² tests, and t-tests were performed to detect significant differences between both cohorts.

RESULTS

The median OS of the overall population was 16 months (95% CI 13.4-23.7) and the median PFS, 2.8 months (95% CI 2.5-3.0). The median OS was longer in cohort B than in cohort A (18.2 vs. 6.1 months; p = 0.071). The best objective response rate to dual ICB was 13.8% and to anti-PD-1 monotherapy 8.9% in the entire population. Patients with liver metastases only had a lower response to dual ICB, yet without significance (cohort A 8.7% vs. cohort B 16.7%; p = 0.45). Adverse events (AE) occurred in 41.6%. Severe AE were observed in 26.3% and evenly distributed between both cohorts.

CONCLUSION

The survival of this large cohort of patients with advanced UM was more favorable than reported in previous benchmark studies. Patients with both hepatic and extrahepatic metastasis showed more favorable survival and higher response to dual ICB than those with hepatic metastasis only.

Abstract CT002: Phase 3 randomized trial comparing tebentafusp with investigator's choice in first line metastatic uveal melanoma

Background: Metastatic uveal melanoma (mUM) has a poor prognosis with a 1-yr OS rate of 52%. No systemic treatment has proven an OS benefit in randomized trials. Tebentafusp (tebe), a bispecific consisting of an affinity-enhanced T cell receptor (TCR) fused to an anti-CD3 effector that can redirect T cells to target gp100+ cells, has shown promising activity in previously treated mUM pts. Here, we report the primary analysis of overall survival (OS) in the intention-to-treat population (ITT) of a Ph3 trial of tebe vs. investigator's choice (IC) as first line (1L) therapy in pts with mUM [NCT03070392].

Materials and Methods: In this randomized, open-label, Ph3 trial, 1L HLA-A*02:01+ pts with mUM were randomized 2:1 to receive tebe or IC of pembrolizumab, ipilimumab or dacarbazine, stratified by LDH. The primary endpoint was OS, defined as the time from randomization to death from any cause. Dual primary objectives were to evaluate 1) OS in the ITT population by comparing all tebe-randomized pts to all IC-randomized pts; and 2) OS in tebentafusp-treated patients with rash during week 1 versus all IC-treated patients. Secondary endpoints included safety and RECIST-defined overall response rate (ORR), progression free survival (PFS) and disease control rate (DCR). Here we present the OS in the ITT population. The study was unblinded by an independent data monitoring committee at the first pre-specified interim analysis. Investigator-reported radiographic-based endpoints were not mature at the first interim analysis. This analysis was conducted on the first interim analysis (data extracted Nov 2020).

Results: 378 pts were randomized to tebe (252) or IC, including pembrolizumab (103), ipilimumab (15) or dacarbazine (7). Tebe significantly prolonged OS compared to IC (HR 0.51; 95% CI 0.36-0.71; P&lt;0.0001) in the ITT population, with estimated 1-yr OS rate of 73.2% (95% CI 66.3-78.9) vs 57.5% (95% CI 47.0-66.6), respectively. The OS benefit of tebe was observed in pre-specified subgroups, including by stratification variable of LDH&gt;ULN and versus pembrolizumab IC. Most common TRAEs were skin-related (gp100+ melanocytes) or cytokine-mediated (T cell activation) and included pyrexia, pruritus, and rash. These AEs decreased in frequency and severity after the first 3-4 doses and were generally manageable with standard interventions. In the tebe arm, the rate of treatment discontinuation due to TRAEs was low (&lt;4%), and there were no treatment-related deaths.

Conclusions: In 1L treatment of mUM pts, tebe monotherapy significantly improved OS compared to IC; the first investigational therapy to improve OS in pts with mUM. Tebe had a predictable and manageable AE profile with a low rate of related discontinuation. Tebe is the first TCR therapeutic to demonstrate an OS benefit.

Citation Format: Sophie Piperno-Neumann, Jessica C. Hassel, Piotr Rutkowski, Jean-Francois Baurain, Marcus O. Butler, Max Schlaak, Ryan J. Sullivan, Sebastian Ochsenreither, Reinhard Dummer, John M. Kirkwood, Anthony M. Joshua, Joseph J. Sacco, Alexander N. Shoushtari, Marlana Orloff, Richard D. Carvajal, Omid Hamid, Shaad E. Abdullah, Chris Holland, Howard Goodall, Paul Nathan. Phase 3 randomized trial comparing tebentafusp with investigator's choice in first line metastatic uveal melanoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr CT002.

Impact of a preceding radiotherapy on the outcome of immune checkpoint inhibition in metastatic melanoma: a multicenter retrospective cohort study of the DeCOG

Background

Immune checkpoint inhibition (ICI) is an essential treatment option in melanoma. Its outcome may be improved by a preceding radiation of metastases. This study aimed to investigate the impact of a preceding radiotherapy on the clinical outcome of ICI treatment.

Methods

This multicenter retrospective cohort study included patients who received anti-cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) or anti-programmed cell death protein 1 (PD-1) ICI with or without preceding radiotherapy for unresectable metastatic melanoma. ICI therapy outcome was measured as best overall response (BOR), progression-free (PFS) and overall survival (OS). Response and survival analyses were adjusted for confounders identified by directed acyclic graphs. Adjusted survival curves were calculated using inverse probability treatment weighting.

Results

835 patients who received ICI (anti-CTLA-4, n=596; anti-PD-1, n=239) at 16 centers were analyzed, whereof 235 received a preceding radiotherapy of metastatic lesions in stage IV disease. The most frequent organ sites irradiated prior to ICI therapy were brain (51.1%), lymph nodes (17.9%) and bone (17.9%). After multivariable adjustment for confounders, no relevant differences in ICI therapy outcome were observed between cohorts with and without preceding radiotherapy. BOR was 8.7% vs 13.0% for anti-CTLA-4 (adjusted relative risk (RR)=1.47; 95% CI=0.81 to 2.65; p=0.20), and 16.5% vs 25.3% for anti-PD-1 (RR=0.93; 95% CI=0.49 to 1.77; p=0.82). Survival probabilities were similar for cohorts with and without preceding radiotherapy, for anti-CTLA-4 (PFS, adjusted HR=1.02, 95% CI=0.86 to 1.25, p=0.74; OS, HR=1.08, 95% CI=0.81 to 1.44, p=0.61) and for anti-PD-1 (PFS, HR=0.84, 95% CI=0.57 to 1.26, p=0.41; OS, HR=0.73, 95% CI=0.43 to 1.25, p=0.26). Patients who received radiation last before ICI (n=137) revealed no better survival than those who had one or more treatment lines between radiation and start of ICI (n=86). In 223 patients with brain metastases, we found no relevant survival differences on ICI with and without preceding radiotherapy.

Conclusions

This study detected no evidence for a relevant favorable impact of a preceding radiotherapy on anti-CTLA-4 or anti-PD-1 ICI treatment outcome in metastatic melanoma.

Clinical characteristics and therapy response in unresectable melanoma patients stage IIIB-IIID with in-transit and satellite metastases

INTRODUCTION

Cutaneous melanoma is notorious for the development of in-transit metastases (ITM). For unknown biological reasons, ITM remain the leading tumour manifestation without progression to distant sites in some patients.

METHODS

In total, 191 patients with initially unresectable stage III ITM and satellite metastases from 16 skin cancer centres were retrospectively evaluated for their tumour characteristics, survival and therapy response. Three groups according to disease kinetics (no distant progress, slow (>6 months) and fast (<6 months) distant progression) were analysed separately.

RESULTS

Median follow-up time was 30.5 (range 0.8-154.0) months from unresectable ITM. Progression to stage IV was observed in 56.5% of cases. Patients without distant metastasis were more often female, older (>70 years) and presented as stage III with lymph node or ITM at initial diagnosis in 45.7% of cases. Melanoma located on the leg had a significantly better overall survival (OS) from time of initial diagnosis compared to non-leg localised primaries (hazard ratio [HR] = 0.61, 95% confidence interval [CI] 0.40-0.91; p = 0.017), but not from diagnosis of unresectable stage III (HR = 0.67, 95% CI 0.45-1.02; p = 0.06). Forty percent of patients received local therapy for satellite and ITM. Overall response rate (ORR) to all local first-line treatments was 38%; disease control rate (DCR) was 49%. In total, 72.3% of patients received systemic therapy for unresectable stage IIIB-D. ORR for targeted therapy (n = 19) was highest with 63.2% and DCR was 84.2% compared to an ORR of 31.4% and a DCR of 54.3% in PD-1 treated patients (n = 70). Patients receiving PD-1 and intralesional talimogene laherparepvec (n = 12) had an ORR of 41.7% and a DCR of 75%.

CONCLUSION

Patients with unresectable ITM and without distant progression are more often female, older, and have a primary on the leg. Response to PD-1 inhibitors in this cohort was lower than expected, but further investigation is required to elucidate the biology of ITM development and the interplay with the immune system.

Overall survival in patients who received checkpoint inhibitors after completing tebentafusp in a phase 3 randomized trial of first-line metastatic uveal melanoma

9526

Background: Tebentafusp (tebe) is a bispecific consisting of an affinity-enhanced T cell receptor fused to an anti-CD3 effector that can redirect T cells to target gp100+ cells. Tebe significantly improved OS compared to investigator’s choice (IC) in first line (1L) mUM [NCT03070392]. In a phase (ph) 2 study of tebe in 2L+ mUM (NCT02570308), several checkpoint inhibitor (CPI) refractory pts who were retreated with CPI after tebe achieved durable clinical benefit [1]. We therefore evaluated clinical outcomes of post-tebe CPI in patients treated on the ph3 trial of tebe versus investigator’s choice (IC) [NCT03070392]. Methods: In the ph3 trial, 378 HLA-A*02:01+ 1L mUM pts were randomized 2:1 to tebe (n=252) or IC (n=126) [pembrolizumab (82%), ipilimumab (12%) or dacarbazine (6%)]. No crossover to tebe was permitted, investigators were free to choose subsequent therapy, and there was no re-randomization at time of subsequent therapy. This analysis was conducted on the first interim analysis (data extracted Nov-2020). When pts received more than one subsequent therapy, the first was used in these analyses. Medians and 1-yr OS from the start of post-study therapy are obtained from standard Kaplan-Meier analyses; hazard ratios (HR) are from Cox regression models adjusted for age and gender. Results: 106/252 (42%) tebe pts received ≥ 1 subsequent therapy: 35% CPI, 9% chemo, 6% liver directed therapy (LDT), 6% other. 55/126 (44%) of IC pts received ≥ 1 subsequent therapy: 21% CPI, 10% chemo, 12% LDT, 10% other. Median time to first subsequent therapy was longer for tebe pts at 5.2 mo vs. IC pts at 3.8 mo. The median duration from start of first subsequent CPI to end date was longer in the prior tebe pts at 4 mo vs prior IC pts at 2.8 mo. From the start of any first subsequent therapy, prior tebe pts had longer OS compared to prior IC pts, HR 0.67 (95% CI 0.42, 1.07). Most of the subsequent therapy was CPI, and the OS benefit was also seen in this subset, HR 0.62 (95% CI 0.34, 1.14). For prior tebe pts, the median and 1-yr OS rates from start of any first subsequent therapy were 13 mo and 53% and from start of first subsequent CPI were 16 mo and 63%. Both were higher than the sequence of IC followed by any therapy (11 mo and 44%), IC followed by CPI (9 mo and 47%) and a recent meta-analysis of 2L+ mUM (7 mo and ̃35% 1-yr OS rate). Conclusions: Pts who progressed on tebe and then received CPI had better OS compared to pts who progressed on IC and then received CPI. Further analysis will explore whether confounding factors are influencing this effect. These exploratory data suggest that tebe, relative to IC, may improve outcomes to subsequent CPI. (1)Yang J. et al. ASCO 2019, J . Cli n Oncol 37:15_suppl, 9592. Clinical trial information: NCT03070392.

Results from the phase Ib of the SENSITIZE trial combining domatinostat with pembrolizumab in advanced melanoma patients refractory to prior checkpoint inhibitor therapy

9545

Background: Anti-PD-1 +/- anti-CTLA4 antibodies are the current standard of care immunotherapy for advanced melanoma. However, a significant proportion of patients do not achieve disease control. Epigenetic modulation, particularly histone deacetylase (HDAC) inhibition, can overcome tumor escape mechanisms and thus might increase the susceptibility to immunotherapy. Methods: Advanced unresectable/metastatic cutaneous melanoma patients primary refractory or non-responding to prior checkpoint inhibitor (CI) therapy were treated with domatinostat at 5 different dose levels (DL) (100 mg (QD), 200 (QD), and 200 mg (BID) using two different schedules (D1-14 and D1-21 q3w) in combination with pembrolizumab (2 mg/kg) q3w to evaluate safety and tolerability. Tumor assessments were performed every 12 weeks and assessed using irRECIST. Sequential tumor biopsies were taken for gene expression analysis and peripheral blood for pharmacokinetic (PK) analysis. Results: We report on preliminary results from the phase Ib part of the ongoing study, data cut-off Feb 1st, 2021 a total of 40 patients have been enrolled. Patient characteristics show that the median number of pretreatments at stage IV was 3, 65 % of patients stage M1c (AJCC 7 or 8) and 35 % with elevated LDH at trial inclusion. Treatment emergent adverse events (AEs) related to domatinostat reported in ≥ 10% of patients were: diarrhea (23%), nausea (20%), fatigue (20%), rash (15%), pyrexia (13%), blood alkaline phosphatase increased (13%), vomiting (10%), dyspnea (10%), all grade 1 and 2 - except one maculo-papular rash grade 3. In total, 8 patients (20 %) developed ≥ grade 3 AEs, with no treatment-related deaths. Patterns of AEs resembled the known safety profiles of domatinostat and pembrolizumab with no increase of immune related AEs for the combination. Maximum tolerated dose was not reached. Four patients discontinued treatment per protocol due to AEs grade 3. We observed clinical activity with 1 complete response, 2 confirmed partial responses and 9 stable diseases (6 confirmed), resulting in a disease control rate of 30% in highly pretreated patients throughout all DLs. Notably, 3 out of 7 patients achieved disease control in DL 3 (domatinostat 200 mg BID D1-14, q3w) and were on treatment ≥ 1.5 years, indicating a trend of dose-dependent clinical activity. Domatinostat treatment resulted in a trend to higher intra-tumoral expression of MHC/APM genes and a more inflamed tumor microenvironment reflecting enhanced T cell infiltration. Conclusions: The combination of domatinostat and pembrolizumab was safe and well tolerated. The observed clinical activity in advanced melanoma patients refractory to previous checkpoint inhibition and the favorable translational findings warrant further development of domatinostat in combination with CI in melanoma and beyond. Clinical trial information: NCT03278665.

Co-primary endpoint of overall survival for tebentafusp (tebe)-induced rash in a phase 3 randomized trial comparing tebe versus investigator’s choice (IC) in first-line metastatic uveal melanoma

9527

Background: Tebe is a bispecific consisting of an affinity-enhanced T cell receptor fused to an anti-CD3 effector that can redirect T cells to target gp100+ cells. In this Phase (Ph) 3, randomized trial of first line (1L) metastatic uveal melanoma (mUM) [NCT03070392], tebe significantly improved overall survival (OS) vs. investigator’s choice (IC) in the intention-to-treat population (ITT). In previous trials, tebe-related skin adverse events (AEs), hypothesized to be on-target, off-tumor activity against gp100-expressing melanocytes, were associated with improved OS. This association was tested prospectively as a co-primary endpoint in the Ph3 study. Methods: 378 1L HLA-A*02:01+ mUM pts were randomized 2:1 to tebe (n = 252) or IC (n = 126). Co-primary endpoints were 1) OS in all randomized pts (ITT) and 2) OS in tebe-randomized pts who develop any grade rash in week (wk) 1 vs. all receiving IC. Rash was defined as composite of preferred AE terms. Melanocyte-related AEs (MRAEs) were defined as pigment change AEs in the skin or hair. Overall study-wide alpha was controlled at 0.05, with 90% assigned to ITT and 10% to rash. This analysis was conducted on the first interim analysis (data extracted Nov-2020). Results: In the 245 tebe treated pts, the characteristic skin related AEs included most frequently rash (at any time) in 201 pts (82%), pruritis in 167 pts (68%), MRAEs in 109 pts (45%) and erythema in 69 pts (28%). While rash, erythema and pruritis mostly occurred in the first 4 weeks, MRAEs occurred after a median of 2.7 mo. Rash captures most pts, 201/227 (89%), who have any of these skin related AEs. Rash occurred in 146 pts (60%) by wk 1; 179 pts (73%) by wk 2; and 195 pts (80%) by wk 3. Tebe pts with wk 1 rash had significantly longer OS vs. the IC arm, HR 0.35 (95% CI 0.23, 0.53), p < 0.0001. The estimated 1-yr OS rates were 83% vs 58%, respectively. When expanded to include tebe pts with rash through wk 3, the 1-yr OS rate of 75% was still numerically higher than IC. The 50 (20%) tebe pts who did not experience rash by week 3 had 1-yr OS rate of 55%. Conclusions: In 1L mUM pts, tebe significantly improved OS compared to IC in the ITT analysis. Week 1 rash, presumed due to tebe redirection of T cells to gp100+ skin melanocytes, was associated with a very strong OS benefit. Therefore, rash may be a marker that the immune system can be mobilized by tebe to target gp100+ cells. The vast majority of tebe pts will develop a rash at some point, and tebe pts without rash may still derive benefit. Clinical trial information: NCT03070392.

Overall survival benefit from tebentafusp in patients with best response of progressive disease

9509

Background: Tebentafusp (tebe) is the first T cell receptor (TCR) therapeutic to demonstrate an overall survival (OS) benefit in a randomized Phase 3 (Ph3) study [ NCT03070392 ]. In Ph2, 42% of pts with best overall response (BOR) of progressive disease (PD) survived > 1 year (yr), suggesting RECIST-based radiographic assessments underestimate OS benefit of tebe. Here we analyzed OS in the Ph3 study in a cohort of pts with BOR of PD by comparing tebe to the control arm of investigator’s choice (IC). Methods: 378 pts were randomized in a 2:1 ratio to tebe vs. IC. BOR was assessed by investigators using RECIST v1.1. Treatment beyond first disease progression (TBP) was permitted for both arms. On the IC arm, only patients receiving pembrolizumab (pembro) continued with TBP and were included in the TBP-related analyses. No crossover to tebe was permitted; investigators were free to choose subsequent therapy. This analysis was conducted on the first interim analysis (data extracted Nov-2020). Kaplan-Meier estimates of OS were based on Day 100 landmark to eliminate immortal time bias and to capture majority of the PDs. Results: By Day 100, PD as BOR occurred in 52% (130/252) of tebe pts (PD-tebe) vs. 60% (76/126) of IC pts (PD-IC). Key baseline characteristics including lactate dehydrogenase, alkaline phosphatase, ECOG performance, age, and sex were similar between PD-tebe vs PD-IC. The proportion of pts with PD due to progression of target lesions (TL), non-TL, or new lesions were also similar between the two groups. More pts received TBP among PD-tebe 53% (69/130) vs PD-pembro 16% (10/61). Median duration of TBP was longer for PD-tebe (7 weeks) vs PD-Pembro (3 weeks). The safety profile of PD-tebe pts during TBP was similar to all tebe-treated pts. OS was superior for PD-tebe vs PD-IC, HR = 0.41 (95%CI 0.25-0.66), even when considering key baseline covariates. While some pts had regression of TL despite diagnosis of PD ( < 10% of pts), the OS benefit remained even when limited to pts with best change of tumor growth of TL, HR 0.46 (0.29, 0.73). 58% (75/130) PD-tebe and 52% (40/76) PD-IC pts received subsequent therapies. In a landmark OS analysis of these pts beginning on 1st day of subsequent therapy, prior tebe was associated with better OS vs. prior IC, HR 0.59 (95%CI 0.36-0.96). Conclusions: Tebe is the first TCR therapeutic to demonstrate an OS benefit in a solid tumor. Surprisingly, a strong OS benefit from tebe is observed even in pts with BOR of PD, suggesting that RECIST-based radiographic assessments do not capture the complete benefit from tebe. The safety profile of tebe during TBP was consistent with that for long-term tebe treatment. Clinical trial information: NCT03070392.

Combining CNN-based histologic whole slide image analysis and patient data to improve skin cancer classification

BACKGROUND

Clinicians and pathologists traditionally use patient data in addition to clinical examination to support their diagnoses.

OBJECTIVES

We investigated whether a combination of histologic whole slides image (WSI) analysis based on convolutional neural networks (CNNs) and commonly available patient data (age, sex and anatomical site of the lesion) in a binary melanoma/nevus classification task could increase the performance compared with CNNs alone.

METHODS

We used 431 WSIs from two different laboratories and analysed the performance of classifiers that used the image or patient data individually or three common fusion techniques. Furthermore, we tested a naive combination of patient data and an image classifier: for cases interpreted as 'uncertain' (CNN output score <0.7), the decision of the CNN was replaced by the decision of the patient data classifier.

RESULTS

The CNN on its own achieved the best performance (mean ± standard deviation of five individual runs) with AUROC of 92.30% ± 0.23% and balanced accuracy of 83.17% ± 0.38%. While the classification performance was not significantly improved in general by any of the tested fusions, naive strategy of replacing the image classifier with the patient data classifier on slides with low output scores improved balanced accuracy to 86.72% ± 0.36%.

CONCLUSION

In most cases, the CNN on its own was so accurate that patient data integration did not provide any benefit. However, incorporating patient data for lesions that were classified by the CNN with low 'confidence' improved balanced accuracy.

Hematological immune related adverse events after treatment with immune checkpoint inhibitors

INTRODUCTION

With the increasing use of checkpoint inhibitors, rare immune-related adverse events (irAE) are being identified. Haematological irAE (hem-irAE) are difficult to treat and have shown high mortality rates. In order to improve side-effect management for these potentially life-threatening events, we analysed frequency, severity and outcomes.

PATIENTS AND METHODS

Patients who developed hem-irAE while being treated with immune checkpoint inhibitors (ICI) therapy were retrospectively identified from 18 international cancer centres.

RESULTS

In total, more than 7626 patients treated with ICI were screened, and 50 patients with hem-irAE identified. The calculated incidence amounts to 0.6% and median onset was 6 weeks after the ICI initiation (range 1-128 weeks). Thrombocytopenia and leucopaenia were the most frequent hem-irAE with 34% (17/50) and 34% (17/50), respectively, followed by anaemia 28% (14/50), hemophagocytic lymphohistiocytosis (4% (2/50)), aplastic anaemia (2% (1/50)), acquired haemophilia A (2% (1/50)) and coagulation deficiency (2% (1/50)). Simultaneous thrombocytopenia and neutropenia occurred in two patients, concurrent anaemia and thrombocytopenia in one patient. Other than cessation of ICI (in 60%) and corticosteroids (in 78%), treatment included second-line immunosuppression in 24% of cases. Events resolved in 78% (39/50), while 18% (9/50) had persistent changes, and 2% (1/50) had fatal outcomes (agranulocytosis).

CONCLUSION

Hem-irAE can affect all haematopoietic blood cell lineages and may persist or even be fatal. Management may require immunosuppression beyond corticosteroids. Although these irAE are rare, treating physicians should be aware, monitor blood counts regularly and promptly act upon detection.

Primary Biliary Cirrhosis and Granulomatous Hepatitis After Immune Checkpoint Blockade in Patients With Metastatic Melanoma: Report of 2 Cases and Literature Discussion

Immune-related adverse events (irAEs) of immune checkpoint inhibitors can potentially affect every organ system, are sometimes challenging, and require a multidisciplinary approach. Most common irAEs are very well characterized, but some other such rare autoimmune liver diseases are probably underdiagnosed and less explored. We present here the case of a 69-year-old man with metastatic melanoma developing a severe primary biliary cirrhosis under pembrolizumab, and of a 52-year-old woman with metastatic melanoma with granulomatous hepatitis in the context of an immune-related multiorgan inflammatory reaction due to ipilimumab and nivolumab. Both cases were in part steroid refractory and required a complex diagnostic assessment and long-term therapeutic management. The liver biopsy was crucial for ensuring a correct diagnosis. Clinicians should be aware of rare liver diseases in the context of increased liver enzymes under immune checkpoint inhibitors, especially if not responding to corticosteroids. The primary diagnostic workup should localize the liver damage (biliary or parenchymal) and distinguish irAEs from other pathologic conditions such as metastasis, second benign and malignant tumors, viral hepatitis, and cholelithiasis. If in doubt, a liver biopsy should be performed. Early diagnosis and accurate assessment of hepatic adverse events is necessary for prompt and effective treatment, with reduction of inappropriate discontinuation of immunotherapy, morbidity, and mortality.

Patterns of care and follow-up care of patients with uveal melanoma in German-speaking countries: a multinational survey of the German Dermatologic Cooperative Oncology Group (DeCOG)

Purpose

Uveal melanoma (UM) is an orphan cancer of high unmet medical need. Current patterns of care and surveillance remain unclear as they are situated in an interdisciplinary setting.

Methods

A questionnaire addressing the patterns of care and surveillance in the management of patients with uveal melanoma was distributed to 70 skin cancer centers in Austria, Germany and Switzerland. Frequency distributions of responses for each item of the questionnaire were calculated.

Results

44 of 70 (62.9%) skin cancer centers completed the questionnaire. Thirty-nine hospitals were located in Germany (88.6%), three in Switzerland (6.8%) and two in Austria (4.5%). The majority (68.2%) represented university hospitals. Most patients with metastatic disease were treated in certified skin cancer centers (70.7%, 29/41). Besides, the majority of patients with UM were referred to the respective skin cancer center by ophthalmologists (87.2%, 34/39). Treatment and organization of follow-up of patients varied across the different centers. 35.1% (14/37) of the centers stated to not perform any screening measures.

Conclusion

Treatment patterns of patients with uveal melanoma in Germany, Austria and Switzerland remain extremely heterogeneous. A guideline for the treatment and surveillance is urgently needed.

Treatment Monitoring of Immunotherapy and Targeted Therapy Using 18F-FET PET in Patients with Melanoma and Lung Cancer Brain Metastases: Initial Experiences

We investigated the value of O-(2-¹⁸F-fluoroethyl)-l-tyrosine (¹⁸F-FET) PET for treatment monitoring of immune checkpoint inhibition (ICI) or targeted therapy (TT) alone or in combination with radiotherapy in patients with brain metastasis (BM) since contrast-enhanced MRI often remains inconclusive. Methods: We retrospectively identified 40 patients with 107 BMs secondary to melanoma (n = 29 with 75 BMs) or non-small cell lung cancer (n = 11 with 32 BMs) treated with ICI or TT who had ¹⁸F-FET PET (n = 60 scans) for treatment monitoring from 2015 to 2019. Most patients (n = 37; 92.5%) had radiotherapy during the course of the disease. In 27 patients, ¹⁸F-FET PET was used to differentiate treatment-related changes from BM relapse after ICI or TT. In 13 patients, ¹⁸F-FET PET was performed for response assessment to ICI or TT using baseline and follow-up scans (median time between scans, 4.2 mo). In all lesions, static and dynamic ¹⁸F-FET PET parameters were obtained (i.e., mean tumor-to-brain ratios [TBR], time-to-peak values). Diagnostic accuracies of PET parameters were evaluated by receiver-operating-characteristic analyses using the clinical follow-up or neuropathologic findings as a reference. Results: A TBR threshold of 1.95 differentiated BM relapse from treatment-related changes with an accuracy of 85% (P = 0.003). Metabolic responders to ICI or TT on ¹⁸F-FET PET had a significantly longer stable follow-up (threshold of TBR reduction relative to baseline, ≥10%; accuracy, 82%; P = 0.004). Furthermore, at follow-up, time to peak in metabolic responders increased significantly (P = 0.019). Conclusion: ¹⁸F-FET PET may add valuable information for treatment monitoring in BM patients treated with ICI or TT.

Addition of Radiotherapy to Immunotherapy: Effects on Outcome of Different Subgroups Using a Propensity Score Matching

Immune checkpoint inhibition (ICI) has been established as successful modality in cancer treatment. Combination concepts are used to optimize treatment outcome, but may also induce higher toxicity rates than monotherapy. Several rationales support the combination of radiotherapy (RT) with ICI as radioimmunotherapy (RIT), but it is still unknown in which clinical situation RIT would be most beneficial. Therefore, we have conducted a retrospective matched-pair analysis of 201 patients with advanced-stage cancers and formed two groups treated with programmed cell death protein 1 (PD-1) inhibitors only (PD1i) or in combination with local RT (RIT) at our center between 2013 and 2017. We collected baseline characteristics, programmed death ligand 1 (PD-L1) status, mutational status, PD-1 inhibitor and RT treatment details, and side effects according to the Common Terminology Criteria for Adverse Events (CTCAE) v.5.0. Patients received pembrolizumab (n = 93) or nivolumab (n = 108), 153 with additional RT. For overall survival (OS) and progression-free survival (PFS), there was no significant difference between both groups. After propensity score matching (PSM), we analyzed 96 patients, 67 with additional and 29 without RT. We matched for different covariates that could have a possible influence on the treatment outcome. The RIT group displayed a trend towards a longer OS until the PD1i group reached a survival plateau. PD-L1-positive patients, smokers, patients with a BMI ≤ 25, and patients without malignant melanoma showed a longer OS when treated with RIT. Our data show that some subgroups may benefit more from RIT than others. Suitable biomarkers as well as the optimal timing and dosage must be established in order to achieve the best effect on cancer treatment outcome.