Abstract
Background:
Despite improvements in the treatment of primary uveal melanoma (UM), patients with metastatic disease continue to exhibit poor survival.
Methods:
We conducted a retrospective review of metastatic UM patients at Yale (initial cohort) and Memorial Sloan Kettering (validation cohort). We utilized Cox proportional hazards regression to determine baseline factors that are associated with overall survival, including sex, Eastern Cooperative Oncology Group (ECOG) Performance Status Scale, laboratory measurements, metastasis location, and use of anti-CTLA-4 and anti-PD-1 therapies. Differences in overall survival were analyzed using Kaplan Meier analysis.
Results:
We identified 89 patients with metastatic UM; 71 and 18, in the initial and validation cohort, respectively. In the initial cohort, median follow-up was 19.8 months (range 2–127 months) and median overall survival was 21.8 months (95% CI 16.6–31.3). Female sex, anti-CTLA-4, and anti-PD-1 therapy were associated with better survival outcomes with adjusted death hazard ratios (HRs) of 0.40 (95% CI 0.20–0.78), 0.44 (0.20–0.97), and 0.42 (0.22–0.84), respectively, whereas development of hepatic metastases and ECOG score ≥ 1 (per 1 U/L) were associated with worse survival outcomes with HRs of 2.86 (1.28–7.13) and 2.84 (1.29–6.09), respectively. In both the initial and validation cohorts, use of immune checkpoint inhibitors was associated with improved overall survival after adjusting for sex and ECOG score, with death HRs of 0.22 (0.08 – 0.56) and 0.04 (0.002 – 0.26), respectively.
Conclusions:
We demonstrate that development of extrahepatic-only metastases, ECOG of 0, immune checkpoint therapy, and female sex were each associated with over two-fold reductions in risk of death.
Keywords: metastatic uveal melanoma, ipilimumab, nivolumab, immunotherapy, tebentafusp, immune checkpoint inhibitors, retrospective cohort study
PLAIN LANGUAGE SUMMARY
Metastatic uveal melanoma patients face limited treatment options and poor survival rates. Results from this retrospective analysis indicate that immune checkpoint inhibitors, such as anti-CTLA-4 and anti-PD-1 therapies, were associated with improved survival outcomes. Factors such as extrahepatic-only metastases, better baseline performance status, and female sex contributed to a more than two-fold reduction in death risk. These findings highlight the potential of immunotherapy in treating metastatic uveal melanoma.
PRECIS
Despite advancements in treatment, metastatic uveal melanoma outcomes remain poor. Female sex, higher baseline performance status, extrahepatic-only metastases, and immune checkpoint therapy are associated with improved survival.
INTRODUCTION
Uveal melanoma (UM) is the most common primary intraocular cancer in adults (1). While treatment modalities for the primary tumor are have excellent local control (>95%), roughly 27–34% of patients develop distant metastases within 10 years (2–5). Once clinical metastases are detected, median overall survival is usually less than 1 year (6).
In January 2021, the US Food and Drug Administration (FDA) approved tebentafusp, a bispecific gp100 peptide-HLA-directed CD3 T cell engager, as a treatment option for patients with metastatic UM (7). Tebentafusp was the first drug to demonstrate a significant impact on overall survival in a large multicenter randomized controlled trial, suggesting a role for immune therapy in the treatment of metastatic uveal melanoma. While use of tebentafusp was associated with significantly better outcomes, its use is limited by its specificity to patients with the HLA-A*02:01 allele, present in only 50% of the population. Therefore, there is still a critical need to identify alternative treatment options for patients who may not be eligible to receive tebentafusp. Previous prospective studies that have explored the role of immune checkpoint inhibitor (ICI) therapies in the treatment of metastatic uveal melanoma (8, 9) have demonstrated modest improvements in overall survival compared to historical rates (10). However, these prospective studies were limited by small sample numbers of approximately 40–50 patients and short follow up periods of approximately 13 months. Here, we leveraged two relatively large clinical cohorts of patients with metastatic uveal melanoma and with long-term follow up, to identify prognostic factors that are associated with improved overall survival.
METHODS
Subjects.
We identified individuals who were treated for metastatic uveal melanoma at Yale New Haven Health (Yale) between March 2007 and November 2021. We reviewed their medical chart and collected patients’ demographic information at first occurrence of metastasis (based on radiographic imaging), including age, sex, race, and ethnicity. We also collected clinical data at first occurrence of metastasis (or within 6 months), including blood lactate dehydrogenase (LDH; in U/L), blood alkaline phosphatase ALP; in U/L), diameter of largest liver metastasis (in mm), and performance status as measured using the Eastern Cooperative Oncology Group (ECOG) scale. If no provider-scores were available, we scored subjects using the ECOG scale as described by Oken et al. using the most contemporaneous clinical notes (11). If no notes were available within the 6-month timeframe but the next subsequent note indicated an ECOG score of 0, an ECOG score of 0 was assumed at baseline. We defined elevated LDH as > 241 U/L and elevated ALP as > 122 U/L. Lastly, we collected dates and locations of all subsequent metastases and also dates of systemic and local treatment modalities used. If tumor diameter was too small to report a precise value, 5 mm was entered per RECIST 1.1.
We also identified a validation cohort of patients with metastatic uveal melanoma from the ocular oncology service at Memorial Sloan Kettering Cancer Center (MSK) between February 2005 and June 2022. Similarly, we reviewed their medical chart and collected patients’ demographic information at first occurrence of metastasis, including age, sex, race, and ethnicity. We also collected baseline ECOG performance status, dates and locations of all metastases and types of systemic and local treatment modalities used.
Outcomes.
The primary outcome was overall survival from time of first metastasis. Dates of death or last follow-up were collected using electronic medical records metadata, provider notes, and records of communications with patients’ next of kin. A secondary outcome was a landmark analysis at 100 days post first-line of immune checkpoint inhibitor therapy, if applicable. We assessed objective response (progressive disease, stable disease, partial response, or complete response) by applying RECIST 1.1 to written imaging reports. If written reports were not available, we assigned a response category based on the clinician’s assessment of response to therapy. Death was considered progression.
Statistical Analysis.
We generated descriptive statistics using percentages for categorical variables and means for continuous variables. Descriptive tables were stratified by treatment regimen: ever having received combination anti-CTLA-4 and anti-PD-1 therapy (“anti-CTLA-4 + anti-PD-1”), ever having received anti-CTLA-4 therapy (“anti-CTLA-4 only”), ever having received anti-PD-1 (“anti-PD-1 only”), or never having received any immune checkpoint inhibitors (“no ICI therapy”). We also generated a swimmer plot to visualize individual-level data of patients and duration of metastatic recurrences. Kaplan-Meier survival analysis was used to visualize overall survival and objective response in different groups and calculate median survival in months. Significance between groups was tested using the Log-rank method. Log-rank test for trend was used to compare more than two groups. Cox proportional hazards regression was used to control for factors found in prior studies to be associated with overall survival, including: sex (12), ECOG (13), LDH (13, 14), ALP (14), and presence of hepatic metastases (13, 15). Subjects missing any variable were excluded from Cox proportional hazards models. Hazard ratios were visualized using forest plots. Statistical analyses and graphs were generated using GraphPad Prism version 9.3.1, (San Diego, California USA) and R version 4.0.4 (R Foundation for Statistical Computing, Vienna, Austria). Swimmer plot was generated using swimplot package in R.
RESULTS
Baseline Characteristics of Initial Cohort
We identified 71 subjects who were diagnosed with metastatic uveal melanoma in the initial cohort. Demographics and clinical characteristics are shown in Table 1. The cohort included 33 females (46.5%). The median age of subjects at time of first metastasis was 66 years (range 16–89 years). The majority (94.4%) of patients identified as White or Caucasian. Subjects’ demographics and clinical characteristics at the time of first metastasis are shown in Table 1. The median time from diagnosis of primary tumor to development of metastatic disease was 34.7 months (range 0–216 months). At the time of their initial diagnosis of metastatic disease, 19 (26.8%) subjects had an ECOG performance score of ≥ 1; an ECOG score could not be assessed for 5 (7.0%) subjects. Hepatic metastases were detected at diagnosis in 56 (78.9%) subjects, and among those with hepatic metastases, the largest lesion measuring 30 mm or greater was present in 20 (35.7%) subjects. At diagnosis of metastatic disease, 22 (31.0%) subjects had an elevated LDH and 11 (15.5%) had an elevated ALP. Median follow-up was 19.8 months (range 2.0–127.1). Clinical outcomes at last follow-up are also shown in Table 1. Median overall survival from time of first metastasis was of 21.8 months (95% CI 16.6–31.3) (Supplementary Figure 1).
Table 1: Demographics of patients diagnosed with metastatic uveal melanoma (initial cohort).
SD = standard deviation; Min=minimum; Max=maximum; ECOG = Eastern Cooperative Oncology; LDH = lactate dehydrogenase (elevated > 241 U/L); ALP = alkaline phosphatase (elevated > 122 U/L); mm = millimeters.
| Anti-CTLA-4 + anti-PD-1 therapy (N=39) |
Anti-CTLA-4 only (N=11) |
Anti-PD-1 only (N=6) |
Other systemic or local therapy (N=6) |
None (N=9) |
Overall (N=71) |
|
|---|---|---|---|---|---|---|
| Baseline | ||||||
| Age (years) | ||||||
| Mean (SD) | 61.1 (14.6) | 58.8 (18.0) | 70.0 (14.0) | 71.2 (10.3) | 78.2 (8.55) | 64.5 (15.3) |
| Median [Min, Max] | 61.8 [20.5, 84.3] | 61.4 [16.2, 88.7] | 69.9 [46.4, 85.3] | 69.0 [55.7, 83.7] | 78.8 [58.2, 86.6] | 66.4 [16.2, 88.7] |
| Sex | ||||||
| Male | 24 (61.5%) | 6 (54.5%) | 3 (50.0%) | 3 (50.0%) | 2 (22.2%) | 38 (53.5%) |
| Female | 15 (38.5%) | 5 (45.5%) | 3 (50.0%) | 3 (50.0%) | 7 (77.8%) | 33 (46.5%) |
| Race | ||||||
| Other | 2 (5.1%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 2 (2.8%) |
| White or Caucasian | 37 (94.9%) | 10 (90.9%) | 5 (83.3%) | 6 (100%) | 9 (100%) | 67 (94.4%) |
| Unknown | 0 (0%) | 1 (9.1%) | 1 (16.7%) | 0 (0%) | 0 (0%) | 2 (2.8%) |
| Ethnicity | ||||||
| Hispanic or Latino | 1 (2.6%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 1 (1.4%) |
| Non-Hispanic | 38 (97.4%) | 9 (81.8%) | 5 (83.3%) | 6 (100%) | 9 (100%) | 67 (94.4%) |
| Unknown | 0 (0%) | 2 (18.2%) | 1 (16.7%) | 0 (0%) | 0 (0%) | 3 (4.2%) |
| Months from Primary to Metastasis | ||||||
| Mean (SD) | 65.8 (59.1) | 30.3 (27.3) | 55.3 (56.0) | 41.2 (75.6) | 48.8 (38.7) | 55.2 (54.7) |
| Median [Min, Max] | 59.1 [0, 216] | 17.6 [6.40, 98.9] | 42.7 [7.47, 154] | 12.8 [0.367, 195] | 50.4 [1.87, 125] | 34.7 [0, 216] |
| ECOG Performance Status | ||||||
| 0 | 31 (79.5%) | 7 (63.6%) | 3 (50.0%) | 5 (83.3%) | 1 (11.1%) | 47 (66.2%) |
| ≥1 | 7 (17.9%) | 1 (9.1%) | 2 (33.3%) | 1 (16.7%) | 8 (88.9%) | 19 (26.8%) |
| Missing | 1 (2.6%) | 3 (27.3%) | 1 (16.7%) | 0 (0%) | 0 (0%) | 5 (7.0%) |
| First Metastatic Lesion | ||||||
| Hepatic | 29 (74.4%) | 7 (63.6%) | 6 (100%) | 6 (100%) | 8 (88.9%) | 56 (78.9%) |
| Mean (SD) | 27.7 (23.6) | 23.3 (11.2) | 22.8 (18.2) | 31.8 (20.7) | 60.6 (53.9) | 31.6 (29.4) |
| Median [Min, Max] | 17.5 [4.50, 96.0] | 23.5 [10.0, 36.0] | 16.0 [7.00, 50.0] | 28.0 [8.00, 56.0] | 80.0 [6.00, 140] | 19.0 [4.50, 140] |
| < 30 | 16 (55.2%) | 4 (57.1%) | 4 (66.7%) | 3 (50.0%) | 3 (37.5%) | 30 (53.6%) |
| ≥ 30 | 10 (34.5%) | 2 (28.6%) | 2 (33.3%) | 2 (33.3%) | 4 (50.0%) | 20 (35.7%) |
| Missing | 3 (10.3%) | 1 (14.3%) | 0 (0%) | 1 (16.7%) | 1 (12.5%) | 6 (10.7%) |
| Extrahepatic | 10 (25.6%) | 4 (36.4%) | 0 (0%) | 0 (0%) | 1 (11.1%) | 15 (21.1%) |
| LDH (U/L) | ||||||
| Mean (SD) | 257 (163) | 188 (44.6) | 240 (54.9) | 340 (137) | 822 (1290) | 311 (449) |
| Median [Min, Max] | 197 [119, 873] | 177 [137, 295] | 251 [163, 312] | 321 [232, 570] | 262 [198, 3450] | 208 [119, 3450] |
| Normal | 19 (48.7%) | 9 (81.8%) | 2 (33.3%) | 2 (33.3%) | 2 (22.2%) | 34 (47.9%) |
| Elevated | 11 (28.2%) | 1 (9.1%) | 3 (50.0%) | 3 (50.0%) | 4 (44.4%) | 22 (31.0%) |
| Missing | 9 (23.1%) | 1 (9.1%) | 1 (16.7%) | 1 (16.7%) | 3 (33.3%) | 15 (21.1%) |
| ALP (U/L) | ||||||
| Mean (SD) | 91.9 (53.0) | 86.3 (23.3) | 87.0 (63.7) | 94.4 (22.4) | 133 (82.3) | 96.2 (53.8) |
| Median [Min, Max] | 75.0 [36.0, 302] | 86.5 [55.0, 127] | 60.0 [41.0, 193] | 91.0 [73.0, 120] | 120 [52.0, 313] | 79.0 [36.0, 313] |
| Normal | 28 (71.8%) | 9 (81.8%) | 4 (66.7%) | 5 (83.3%) | 4 (44.4%) | 50 (70.4%) |
| Elevated | 5 (12.8%) | 1 (9.1%) | 1 (16.7%) | 0 (0%) | 4 (44.4%) | 11 (15.5%) |
| Missing | 6 (15.4%) | 1 (9.1%) | 1 (16.7%) | 1 (16.7%) | 1 (11.1%) | 10 (14.1%) |
| Last Follow-Up | ||||||
| Disposition | ||||||
| Alive | 8 (20.5%) | 0 (0%) | 0 (0%) | 1 (16.7%) | 1 (11.1%) | 10 (14.1%) |
| Deceased | 31 (79.5%) | 11 (100%) | 6 (100%) | 5 (83.3%) | 8 (88.9%) | 61 (85.9%) |
Cox Proportional Hazards Analysis of Overall Survival
We analyzed the association of sex, baseline clinical variables (presence of hepatic metastases, ECOG score, laboratory values (ALP and LDH)), and treatment variables (overall use of anti-CTLA-4 and anti-PD-1 therapies) with overall survival using multivariable cox proportional hazards regression analysis. Female sex, anti-CTLA-4, and anti-PD-1 therapies were associated with improved survival outcomes with death HRs of 0.40 (0.20–0.78; P = 0.009), 0.44 (95% CI 0.20–0.97; P = 0.037), and 0.42 (0.22–0.84; P = 0.013), respectively, whereas presence of hepatic (vs. extrahepatic-only) metastases or an ECOG score of ≥ 1 were associated with worse overall survivals with death HRs of 2.86 (1.28–7.13; P = 0.016), and 2.84 (1.29–6.09; P = 0.008), respectively (Figure 1A) Elevated levels of LDH or ALP were not significantly associated with survival (P = 0.95). In an alternative model, we included the above variables but stratified subjects based on the treatment regimen: those receiving any ICI agents, those receiving local or systemic therapy other than ICI, and those who did not receive any treatment. In this model, female sex and use of ICI therapy (vs. other systemic and local therapy) were associated with improved survival outcomes, with death HRs of 0.43 (0.22–0.83; P = 0.013), and 0.16 (0.05–0.62; P = 0.003), respectively, whereas presence of hepatic metastases and ECOG ≥ 1 were associated with worse overall survival, with death HRs of 2.59 (1.18–6.40; P = 0.026), and 2.91 (1.15–6.81; P = 0.017), respectively. No treatment intervention and elevated levels of LDH or ALP were not significantly associated with survival (P = 0.48 and 0.87, respectively) (Figure 1B).
Figure 1: Multivariable Cox proportional hazard analysis for overall survival.
Each point represents the hazard ratio (HR) with 95% confidence intervals (CI) for the corresponding value of the given covariate compared to its reference. Covariates in model A included hepatic (vs. extrahepatic only) metastases at baseline, ECOG ≥ 1 (vs. 0), elevated LDH or ALP, use of anti-CTLA-4, use of anti-PD-1, and female sex; B replaced anti-CTLA-4 and anti-PD-1 with a single treatment variable (ICI vs. other local/systemic treatment vs. no treatment). Of 71 patients, 50 experienced events (death) and 12 were excluded due to missing data.
Kaplan Meier Analyses of Overall Survival
Next, we analyzed overall survival of subjects stratified by each variable and calculated median overall survival using Kaplan Meier survival analysis. Baseline hepatic metastases (median OS 19.3 months, 95% CI 12.8–29.2) was associated with lower overall survival compared to patients who presented with extra-hepatic metastases only (35.8, 22.4-NA; P = 0.006) (Figure 2A). An ECOG score of ≥ 1 was associated with decreased survival (12.2, 7.8–22.4) compared to an ECOG score of 0 (30.2, 20.3–35.2; P = 0.001) (Figure 2B). Anti-CTLA-4 use was associated with improved survival (30.3, 21.8–35.8) compared to none (8.9, 6.1–24.8; P<0.001) (Figure 2C), and anti-PD-1 therapy use (28.4, 17.4–35.2) was significantly associated with improved survival over none (16.5, 6.4–29.2; P = 0.011) (Figure 2D). Female sex (24.8, 12.1–35.8) was not significantly associated with improved survival compared to male sex (21.8, 14.3–31.3; P = 0.507) (Figure 2E). Elevated LDH levels were not significantly associated with worse survival (31.3, 20.3–35.8) compared normal LDH levels (13.3, 8.9–30.2; P = 0.061) (Figure 2F). Elevated ALP levels were not significantly associated with worse survival (22.9, 17.4–32.2) compared to normal ALP levels (6.8, 3.9-NA; P=0.082) (Figure 2G). Presence of a hepatic lesion ≥ 30 mm in diameter (8.9, 6.2–28.4) was associated with worse survival compared to presence of a hepatic lesion < 30 mm in diameter (25.4, 19.3–35.2; P = 0.003) (Supplementary Figure 2).
Figure 2: Probability of overall survival based on baseline presence or absence of hepatic metastases, performance status, treatment agents, sex, and laboratory values.
Kaplan-Meier survival curves for subjects stratified based on A) baseline presence of hepatic (vs. extrahepatic-only) metastases, B) baseline ECOG performance score, C) use of anti-CTLA4 therapy, D) use of anti-PD-1 therapy, E) sex, F) LDH, and G) ALP. Numbers of subjects and median overall survival with 95% CI are shown for each subgroup. Statistical significance tested using two-sided log-rank test.
Determinants of survival in patients with hepatic metastasis
Given that presence of hepatic metastasis at baseline was associated with worse survival we analyzed survival at the individual level data to rule out the possibility that improved survival among those with extrahepatic metastases could have been driven by few long-term survivors. The overall survival times from diagnosis of metastatic disease until death or last follow up are presented using swimmer plots, wherein subjects were ranked based on the duration since metastatic disease onset. Subjects without hepatic metastases clustered separately and displayed longer overall survival times (Figure 3). We then sought to analyze determinants of poor survival among those with hepatic metastases at baseline. We analyzed the association of ECOG, ICI therapy, elevated baseline laboratory measurements (LDH or ALP), and long diameter of the largest hepatic lesion at first metastasis using cox-proportional hazard regression analysis. Larger metastatic lesion diameter (1.03, 1.02–1.05; P < 0.001) and ECOG score of ≥ 1 (5.26, 1.63–15.6; P = 0.003) were associated with significantly increased death HR, whereas use of ICI therapy was associated with significantly decreased death HR (0.14, 0.04–0.59; P = 0.003) compared to use of only other local or systemic treatments (Supplementary Figure 3). Elevated baseline LDH or ALP measurements (P = 0.616) and the absence of active treatment (P = 0.229) were not significantly associated with survival.
Figure 3: Swimmer plot of months since first metastasis.
The left-most border of each column segment is the point in time at which new sites of metastasis were discovered. Color indicates whether new metastases were hepatic (red) or extrahepatic (grey). Arrows indicate patient alive recent follow-up. The top panel (A) includes deceased patients and the bottom panel (B) patients alive at last follow-up.
Association of immune checkpoint inhibitor therapy with overall survival
Given that use of anti-CTLA-4 or anti-PD-1 agents was associated with better outcomes, we analyzed the effect of combination therapy on survival. Of the 56 subjects who received any ICI, 39 (69.6%) received both anti-CTLA-4 and anti-PD-1 agents, 11 (19.6%) only received anti-CTLA-4 therapy, and 6 (10.7%) only received anti-PD-1 therapy, respectively (Table 1). Use of any immune checkpoint inhibitor was associated with improved survival (29.2, 20.9–35.2) compared to no use (6.3, 5.8-NA; P<0.001) (Supplementary Figure 4A). Use of combination anti-CTLA-4/anti-PD-1 therapy (30.3, 21.8–50.3) was not associated with better survival compared to anti-CTLA-4 only therapy (29.2, 20.3-NA; P = 0.323) but was associated with better survival compared to anti-PD-1 therapy (10.6, 8.9-NA; P = 0.033) or no ICI therapy (6.3, 5.8-NA; P < 0.001). Use of anti-CTLA-4 only was associated with better survival than anti-PD-1 only (P = 0.031) and no ICI therapy (P < 0.001). Anti-PD-1 only therapy (10.6, 8.9-NA) was not significantly associated with greater survival compared to no ICI (P = 0.22) (Supplementary Figure 4B).
Survival outcomes comparing those who received ICI and those who did not, stratified by different variables, are shown in Supplementary Figure 5. Use of ICI therapy was associated with improved OS among subjects with an ECOG of 0 (P < 0.001), female sex (P = 0.001), male sex (P = 0.001), patients with normal LDH (P = 0.022), elevated LDH (P = 0.031), normal ALP (P = 0.003), elevated ALP (P = 0.002), and hepatic metastases (P < 0.001). ICI therapy was not associated with improved survival among patients with ECOG ≥ 1 (P = 0.476). There was only one subject with extrahepatic-only metastases that received ICI; therefore, a comparison in that subgroup was not made.
We examined the 100-day response to the first line of immune checkpoint inhibitor (ICI)-containing regimen that patients received. Among 56 patients who received ICI therapy, 21 (37.5%) achieved stable disease or partial response (SD/PR) and 32 (57.1%) experienced progressive disease (PD); records were not available during the post-landmark period for 3 patients (Supplementary Table 1). The 100-day response to ICI therapy was predictive of overall survival. Median OS (95% CI) among patients who exhibited SD/PR within the first 100 days of initiating immune checkpoint inhibitor therapy was 33.8 months (22.6-NA) compared to 4.9 months (3.8–9.6) among patients who exhibited PD (Figure 4A). There was also a trend toward overall survival among patients who had PD if they received anti-CTLA4 therapy. Medial OS for those with PD who received a regimen containing CTLA-4 inhibitor was 5.7 (4.0–18.8), compared to 2.8 (1.8-NA) for those who did not, P = 0.054 (Figure 4B).
Figure 4: Overall survival calculated from landmark.
Stratified by response (stable disease or partial response (SD/PR) vs. progressive disease (PD) to first course of ICI regimen (A). Among patients with PD, stratified by type of ICI (B).
Female sex and survival
Given that female sex was significantly associated with better overall survival in the multivariable model but not in the univariate Kaplan Meier analysis, we examined the distribution of covariates between sexes and found no significant differences in rates of baseline ECOG of 0 between both sexes (males, 73.7% vs. females, 57.6%; P = 0.309), elevated LDH (34.2% vs. 27.3%; P = 0.695), elevated ALP (13.2% vs. 18.2%; P = 0.857), baseline hepatic metastases (81.6% vs. 75.8%; P = 0.758), use of anti-CTLA-4 therapy (78.9% vs. 60.6%; P= 0.153), or use of anti-PD-1 therapy (71.1% vs. 54.5%; P = 0.233) (Supplementary Table 2). Among patients who had hepatic metastasis at initial presentation, there was no significant difference in the mean diameter of the largest lesion between both sexes (28.5 mm vs. 35.3 mm; P = 0.442). To identify covariates that modified the effect of female sex on overall survival, we applied stepwise backwards selection on the Cox proportional hazards regression analysis for overall survival. Different model iterations are included in Supplementary Table 3. We found that, at minimum, controlling for ECOG performance status, initial site of metastasis, and anti-PD-1 use was required for female sex to be significantly associated with overall survival in the multivariable analysis.
Validation Cohort
The validation cohort consisted of 18 patients. Mean age was 63 years (SD 15 years). There were 8 (44.4%) female patients, 16 (88.9%) identified as White, 16 (88.9%) identified as non-Hispanic, and 17 (94.4%) presented initially with hepatic metastases (Supplementary Table 4). At last follow-up 15 (83.3%) patients were deceased with a median follow-up of 13.4 months (range 0.5–33.3 months). Overall survival was 17.6 months (95% CI 11.2–31.6). We then sought to determine whether ICI use was associated with significantly improved overall survival. Given that only few subjects had extrahepatic metastases in the validation cohort, we analyzed survival only in those with hepatic metastases, in both the initial (n = 56) and validation (n = 17) cohorts; 6 patients from the initial cohort and 1 patient from the validation cohort were removed due to missing data. In two Cox proportional hazards regression models including ICI use, ECOG status and sex, ICI use was associated with improved overall survival in both the initial and validation cohorts, with death HRs of 0.22 (0.08 – 0.56; P = 0.002) and 0.04 (0.002 – 0.257; P = 0.004), respectively (Figure 5A). Estimated survival of subjects stratified by ICI use, after adjusting for ECOG status and sex, demonstrated improved survival in both cohorts (Figure 5B).
Figure 5: Cox proportional hazards adjusted survival of patients with hepatic disease at initial diagnosis of metastatic uveal melanoma stratified by use of ICI therapy.
Separate Cox proportional hazards models adjusting for sex, baseline ECOG (0 vs. ≥ 1), and any receipt of ICI therapy (yes vs. no) were created by cohort (initial and validation). The hazard ratios for ICI use (vs. no use) with 95% confidence intervals are plotted by cohort in panel A. The adjusted cumulative survival curves stratified by receipt of ICI are plotted in panel B.
DISCUSSION
In this relatively large cohort of patients diagnosed with metastatic uveal melanoma, and with long-term follow up data, we identified factors associated with overall survival. In multivariate analyses, use of immune checkpoint inhibitor therapy and female sex were associated with increased overall survival, whereas poorer performance status and hepatic metastases were associated with decreased overall survival.
We observed better overall survival among patients who were treated with immune checkpoint inhibitor therapy compared to those who were not. Use of CTLA-4 inhibitors, with or without PD-1 inhibitors, in the setting of metastatic UM has been examined in previous studies including two phase II studies and three retrospective analyses (Supplementary Table 5) (8, 9, 13, 16, 17). The two single-arm phase II studies demonstrated a modest survival benefit with median overall survival rates of 13–19 months (8, 9), compared to a historical benchmark of 9 months for medical treatment (10). Retrospective analyses of immune checkpoint inhibitors have observed a median overall survival of 7–16 months (13, 16–18). Salaün et al. reported a low response rate (4%) in their retrospective analysis of 47 patients treated with combination anti-CTLA-4/anti-PD-1 therapy that was comparable with monotherapy (18–20). In our cohort, there was a survival benefit for individuals who received immune checkpoint inhibitors in the univariate and multivariate analyses after adjusting for confounding baseline variables.
Several factors may have contributed to the discrepancy between our findings and prior observations. The current study has the longest reported follow up intervals, with a median follow up of 19.8 months (range 2.0 to 127.1 months) in the initial cohort. Our cohorts were comprised of 89 patients, which is a relatively large number of metastatic uveal melanoma patients and may have contributed to achieving statistical significance compared to the other cohorts where a trend toward improved survival was demonstrated but did not reach statistical significance. We also identified significant factors that may impact survival such as presence of hepatic metastasis at baseline, sex, and performance status. Differences in distribution of these variables may account for differences in overall survival between the different studies. We found that the initial 100-day response to immune checkpoint inhibitor therapy was predictive of overall survival, an observation that is line with a retrospective analysis by Koch et al., where patients who had an initial partial or complete response to immune checkpoint inhibitor therapy demonstrated improved overall survival (13).
Tebentafusp is the first immune therapy agent that has demonstrated a significant impact on overall survival in a randomized controlled trial (7). This has renewed interest in immune therapy as a therapeutic option for metastatic uveal melanoma. Tebentafusp’s use is limited by its specificity to patients with a specific HLA-A*02:01 allele. Therefore, there is still a critical need to identify other treatment options. In the tebentafusp’s phase III trial, patients receiving tebentafusp had median overall survival of 21.7 months (95% CI 18.6 – 28.6 months), which was significantly higher than a median overall survival of 16 months (9.7 – 18.4) in the control arm, which consisted of the investigator’s choice of either pembrolizumab (PD-1 inhibitor), ipilimumab (CTLA-4 inhibitor), or dacarbazine (7). In the subgroup analysis, survival outcomes for patients receiving tebentafusp were superior to those who received PD-1 inhibitor but were equal to those who received CTLA-4 inhibitors (HR 0.89, 95% CI 0.38–2.31). In our study, those who received any immune checkpoint inhibitor had a median overall survival of 29.2 months (20.9–35.2 months). Although comparison of single-agent regimens is limited by small cohort numbers, use of anti-CTLA-4 only was associated with better survival than anti-PD-1 only (P = 0.03). In addition, there was a trend toward higher survival among those with progressive disease who received anti-CTLA4 therapy compared to those who had progressive disease but did not receive anti-CTLA4 therapy. However, given that the majority (89%) of subjects who received immune checkpoint inhibitors in our cohorts received anti-CTLA-4 therapy, we cannot rigorously compare the efficacy of anti-PD-1 and anti-CTLA-4 therapy. Additionally, while the survival rates of those who received immune checkpoint inhibitors in our cohort is comparable to the survival outcomes reported in the tebentafusp trial, we cannot directly compare outcomes between the different agents and across different studies. Nevertheless, our data, along with that from the subgroup analysis in the tebentafusp study, suggests that a large randomized controlled trial comparing anti-CTLA-4 therapy to tebentafusp, ideally to patients with the HLA-A*02:01 haplotype, may be warranted.
We found that baseline presence of hepatic metastases was associated with worse survival outcomes, and among those with hepatic metastases, increased tumor diameter was associated with lower overall survival. The significance of metastatic uveal melanoma to the liver as a poor prognostic factor was first demonstrated by Nicholas et al., where presence of hepatic metastases was associated with decreased overall survival (HR 2.81, 95% CI 1.30–6.08) (15). However, this association was not significant when adjusting for other covariates in the multivariable Cox regression analysis. Similarly, these trends were observed in two phase II studies of anti-CTLA-4 and anti-PD-1 combination therapy, albeit not reaching statistical significance (8, 9). Pelster et al. demonstrated a trend toward increased progression-free survival (PFS) among patients with extrahepatic metastases compared to those with hepatic metastases but did not demonstrate a statistically significant improvement in either progression free survival or overall survival (8). Similarly, Piulats et al. demonstrated a trend toward decreased survival among patients without extrahepatic metastases (compared to those with extrahepatic metastases) (9). Koch et al. found that hepatic metastases were more common among patients who survived < 2 years compared to patients who survived longer (13). It is possible that the relatively large number of subjects in our cohort, compared to other studies, may have contributed to achieving statistical significance. Our results, and these recent trends demonstrating difference in outcomes based on metastases location, contrast with earlier studies. Rantala et al. analyzed determinants of long-term survival in a Finnish nationwide cohort of 330 patients diagnosed with metastatic choroidal and ciliary body melanoma between 1999–2016. Rates of hepatic-only metastatic disease were not significantly different among short-term (< 25 months), intermediate-term (25–42 months) and long-term (≥ 42 months) survival cohorts. A large metanalysis by Khoja et al. of 912 patients from 29 studies found no difference in survival between hepatic and extrahepatic metastasis (10). However, the majority of subjects analyzed in these older studies did not receive immune therapy. It is possible that differences in survival rates between those with hepatic metastases and those with extrahepatic-only metastases may be mediated by the use of immune checkpoint inhibitors. We could not test this hypothesis in this study since there was only one patient in our cohort who had extrahepatic metastases (without hepatic involvement) and did not receive immune therapy.
We also found that female sex was associated with better overall survival in the multivariate model but not in univariate analysis of our initial cohort. We found that adjusting for ECOG performance status, initial site of metastasis, and anti-PD-1 therapy were necessary to demonstrate a significant relationship between sex and overall survival. Prior studies also suggest that response to immune checkpoint inhibitor therapy in other cancer subtypes may vary by sex, although findings are mixed (12, 21).
Several limitations should be considered when interpreting our study’s findings. First, our data come from two academic centers located in high-income US states. Early access to care may lead to lead-time bias and improved survival rates. Second, this analysis is based on retrospective chart review and patients were not randomized to different treatment modalities. While we adjusted for ECOG score, LDH, and ALP at baseline, the treatment decision may have been affected by selection bias. Immune therapies are associated with potentially severe adverse effects (7–9). Clinicians may have been more inclined to recommend ICI, especially combination regimens, to patients who were healthier at baseline. We observed that 80% of patients who received combination anti-CTLA-4/anti-PD-1 therapy had a baseline ECOG performance status of 0, compared to 64% among those who received anti-CTLA-4 alone, 50% among those who received anti-PD-1 alone, 83% among those who received another systemic or local therapy, and 11% among those who did not receive active treatment. Third, many patients received combination therapy. Nonetheless, this represents a relatively large cohort for a rare metastatic disease for which few treatment options exist.
In summary, we found that the absence of hepatic metastases, lower LDH, increased performance status, treatment with anti-CTLA-4 or anti-PD-1 therapy, and female sex were associated with improved overall survival in patients diagnosed with metastatic uveal melanoma. Our findings suggest that immune checkpoint inhibitors may provide a viable treatment option for patients with metastatic uveal melanoma.
Data Availability:
The source data are available upon reasonable request to qualified researchers.
Supplementary Material
Acknowledgements
The authors would like to thank Melissa Robbins, M.P.H., Raymond Baumann Jr, Ph.D., and the Melanoma Registry at Yale University.
Funding:
This work was supported by a grant from the National Cancer Institute (NIH), Research Grant P30 CA016359 (MFB) and by a grant from the Connecticut Lions Eye Research Foundation (MFB). CAB, ANS, JHF, and DHA receive support from the Cancer Center Support Grant P30 CA008748 National Cancer Institute (NIH). SM is supported by the National Institute on Aging of the National Institutes of Health under Award Number T35AG049685. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Footnotes
Conflict of Interest: The authors have no conflicts of interest to declare.
Ethics Approval: The study adhered to the tenets of Helsinki and was approved by the Yale Institutional Review Board (IRB) (#2000031254) and MSK IRB (#16-1464).
Patient Consent: A waiver of informed consent was obtained from the IRB for retrospective chart review.
REFERENCES
- 1.Singh AD, Turell ME, Topham AK. Uveal melanoma: trends in incidence, treatment, and survival. Ophthalmology 2011;118(9):1881–5. Epub 2011/06/28. doi: 10.1016/j.ophtha.2011.01.040. [DOI] [PubMed] [Google Scholar]
- 2.Shields CL, Furuta M, Thangappan A, Nagori S, Mashayekhi A, Lally DR, Kelly CC, Rudich DS, Nagori AV, Wakade OA, Mehta S, Forte L, Long A, Dellacava EF, Kaplan B, Shields JA. Metastasis of Uveal Melanoma Millimeter-by-Millimeter in 8033 Consecutive Eyes. Archives of Ophthalmology 2009;127(8):989–98. doi: 10.1001/archophthalmol.2009.208. [DOI] [PubMed] [Google Scholar]
- 3.Radivoyevitch T, Zabor EC, Singh AD. Uveal melanoma: Long-term survival. PLOS ONE 2021;16(5):e0250939. doi: 10.1371/journal.pone.0250939. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Diener-West M, Reynolds SM, Agugliaro DJ, Caldwell R, Cumming K, Earle JD, Hawkins BS, Hayman JA, Jaiyesimi I, Jampol LM, Kirkwood JM, Koh W-J, Robertson DM, Shaw JM, Straatsma BR, Thoma J, Collaborative Ocular Melanoma Study G. Development of metastatic disease after enrollment in the COMS trials for treatment of choroidal melanoma: Collaborative Ocular Melanoma Study Group Report No. 26. Archives of ophthalmology (Chicago, Ill : 1960) 2005;123(12):1639–43. doi: 10.1001/archopht.123.12.1639. [DOI] [PubMed] [Google Scholar]
- 5.Diener-West M, Reynolds SM, Agugliaro DJ, Caldwell R, Cumming K, Earle JD, Green DL, Hawkins BS, Hayman J, Jaiyesimi I, Kirkwood JM, Koh W-J, Robertson DM, Shaw JM, Thoma J. Screening for Metastasis From Choroidal Melanoma: The Collaborative Ocular Melanoma Study Group Report 23. Journal of Clinical Oncology 2004;22(12):2438–44. doi: 10.1200/JCO.2004.08.194. [DOI] [PubMed] [Google Scholar]
- 6.Rantala ES, Hernberg M, Kivelä TT. Overall survival after treatment for metastatic uveal melanoma: a systematic review and meta-analysis. Melanoma research 2019;29(6):561–8. doi: 10.1097/CMR.0000000000000575. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Nathan P, Hassel JC, Rutkowski P, Baurain J-F, Butler MO, Schlaak M, Sullivan RJ, Ochsenreither S, Dummer R, Kirkwood JM, Joshua AM, Sacco JJ, Shoushtari AN, Orloff M, Piulats JM, Milhem M, Salama AKS, Curti B, Demidov L, Gastaud L, Mauch C, Yushak M, Carvajal RD, Hamid O, Abdullah SE, Holland C, Goodall H, Piperno-Neumann S. Overall Survival Benefit with Tebentafusp in Metastatic Uveal Melanoma. N Engl J Med 2021;385(13):1196–206. doi: 10.1056/NEJMoa2103485. [DOI] [PubMed] [Google Scholar]
- 8.Pelster MS, Gruschkus SK, Bassett R, Gombos DS, Shephard M, Posada L, Glover MS, Simien R, Diab A, Hwu P, Carter BW, Patel SP. Nivolumab and Ipilimumab in Metastatic Uveal Melanoma: Results From a Single-Arm Phase II Study. Journal of Clinical Oncology 2020;39(6):599–607. doi: 10.1200/JCO.20.00605. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Piulats JM, Espinosa E, de la Cruz Merino L, Varela M, Alonso Carrión L, Martín-Algarra S, López Castro R, Curiel T, Rodríguez-Abreu D, Redrado M, Gomà M, Rullán AJ, Calvo González A, Berrocal-Jaime A. Nivolumab Plus Ipilimumab for Treatment-Naïve Metastatic Uveal Melanoma: An Open-Label, Multicenter, Phase II Trial by the Spanish Multidisciplinary Melanoma Group (GEM-1402). Journal of Clinical Oncology 2021;39(6):586–98. doi: 10.1200/JCO.20.00550. [DOI] [PubMed] [Google Scholar]
- 10.Khoja L, Atenafu EG, Suciu S, Leyvraz S, Sato T, Marshall E, Keilholz U, Zimmer L, Patel SP, Piperno-Neumann S, Piulats J, Kivelä TT, Pfoehler C, Bhatia S, Huppert P, Van Iersel LBJ, De Vries IJM, Penel N, Vogl T, Cheng T, Fiorentini G, Mouriaux F, Tarhini A, Patel PM, Carvajal R, Joshua AM. Meta-analysis in metastatic uveal melanoma to determine progression free and overall survival benchmarks: an international rare cancers initiative (IRCI) ocular melanoma study. Annals of Oncology 2019;30(8):1370–80. doi: 10.1093/annonc/mdz176. [DOI] [PubMed] [Google Scholar]
- 11.Oken MM, Creech RH, Tormey DC, Horton J, Davis TE, McFadden ET, Carbone PP. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol 1982;5(6):649–55. Epub 1982/12/01. [PubMed] [Google Scholar]
- 12.Conforti F, Pala L, Bagnardi V, De Pas T, Martinetti M, Viale G, Gelber RD, Goldhirsch A. Cancer immunotherapy efficacy and patients’ sex: a systematic review and meta-analysis. The Lancet Oncology 2018;19(6):737–46. doi: 10.1016/S1470-2045(18)30261-4. [DOI] [PubMed] [Google Scholar]
- 13.Koch EAT, Petzold A, Wessely A, Dippel E, Erdmann M, Heinzerling L, Hohberger B, Knorr H, Leiter U, Meier F, Mohr P, Rahimi F, Schell B, Schlaak M, Terheyden P, Schuler-Thurner B, Ugurel S, Utikal J, Vera J, Weichenthal M, Ziller F, Berking C, Heppt MV. Clinical determinants of long-term survival in metastatic uveal melanoma. Cancer Immunology, Immunotherapy 2022;71(6):1467–77. doi: 10.1007/s00262-021-03090-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Rantala ES, Parrozzani R, Hernberg MM, Chiarion-Sileni V, Kivelä TT, Midena E. Determinants of Long-Term Survival in Metastatic Choroidal and Ciliary Body Melanoma: Long-Term Survival in Metastatic Ciliochoroidal Melanoma. Am J Ophthalmol 2022. Epub 2022/11/04. doi: 10.1016/j.ajo.2022.10.017. [DOI] [PubMed] [Google Scholar]
- 15.Nicholas MN, Khoja L, Atenafu EG, Hogg D, Quirt I, Butler M, Joshua AM. Prognostic factors for first-line therapy and overall survival of metastatic uveal melanoma: The Princess Margaret Cancer Centre experience. Melanoma Res 2018;28(6):571–7. Epub 2018/08/02. doi: 10.1097/cmr.0000000000000468. [DOI] [PubMed] [Google Scholar]
- 16.Heppt MV, Amaral T, Kähler KC, Heinzerling L, Hassel JC, Meissner M, Kreuzberg N, Loquai C, Reinhardt L, Utikal J, Dabrowski E, Gesierich A, Pföhler C, Terheyden P, Thoms K-M, Zimmer L, Eigentler TK, Kirchberger MC, Stege HM, Meier F, Schlaak M, Berking C. Combined immune checkpoint blockade for metastatic uveal melanoma: a retrospective, multi-center study. Journal for ImmunoTherapy of Cancer 2019;7(1):299. doi: 10.1186/s40425-019-0800-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Najjar YG, Navrazhina K, Ding F, Bhatia R, Tsai K, Abbate K, Durden B, Eroglu Z, Bhatia S, Park S, Chowdhary A, Chandra S, Kennedy J, Puzanov I, Ernstoff M, Vachhani P, Drabick J, Singh A, Xu T, Yang J, Carvajal R, Manson D, Kirkwood JM, Cohen J, Sullivan R, Johnson D, Funchain P, Shoushtari A. Ipilimumab plus nivolumab for patients with metastatic uveal melanoma: a multicenter, retrospective study. J Immunother Cancer 2020;8(1). Epub 2020/06/26. doi: 10.1136/jitc-2019-000331. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Salaün H, de Koning L, Saint-Ghislain M, Servois V, Ramtohul T, Garcia A, Matet A, Cassoux N, Mariani P, Piperno-Neumann S, Rodrigues M. Nivolumab plus ipilimumab in metastatic uveal melanoma: a real-life, retrospective cohort of 47 patients. Oncoimmunology 2022;11(1):2116845. Epub 2022/09/13. doi: 10.1080/2162402x.2022.2116845. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Hodi FS, Chiarion-Sileni V, Gonzalez R, Grob JJ, Rutkowski P, Cowey CL, Lao CD, Schadendorf D, Wagstaff J, Dummer R, Ferrucci PF, Smylie M, Hill A, Hogg D, Marquez-Rodas I, Jiang J, Rizzo J, Larkin J, Wolchok JD. Nivolumab plus ipilimumab or nivolumab alone versus ipilimumab alone in advanced melanoma (CheckMate 067): 4-year outcomes of a multicentre, randomised, phase 3 trial. Lancet Oncol 2018;19(11):1480–92. Epub 2018/10/27. doi: 10.1016/s1470-2045(18)30700-9. [DOI] [PubMed] [Google Scholar]
- 20.Rossi E, Pagliara MM, Orteschi D, Dosa T, Sammarco MG, Caputo CG, Petrone G, Rindi G, Zollino M, Blasi MA, Cassano A, Bria E, Tortora G, Schinzari G. Pembrolizumab as first-line treatment for metastatic uveal melanoma. Cancer Immunol Immunother 2019;68(7):1179–85. Epub 2019/06/09. doi: 10.1007/s00262-019-02352-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Wallis CJD, Butaney M, Satkunasivam R, Freedland SJ, Patel SP, Hamid O, Pal SK, Klaassen Z. Association of Patient Sex With Efficacy of Immune Checkpoint Inhibitors and Overall Survival in Advanced Cancers: A Systematic Review and Meta-analysis. JAMA Oncology 2019;5(4):529–36. doi: 10.1001/jamaoncol.2018.5904. [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
Data Availability Statement
The source data are available upon reasonable request to qualified researchers.