Abstract
Background
One in every 41 women develops malignant melanoma in their lifetime, with noncutaneous melanomas arising in areas such as the genitourinary (GU) system being particularly rare and aggressive. We used machine learning (ML) to build prognostic models for vaginal (VaM) and vulvar (VuM) melanomas and developed the first predictive web-based tool for survival in these cancers.
Methods
We leveraged the SEER database (2000–2020) to assemble our cohort and extract relevant clinical and demographic variables. Prognostic factors were screened using univariate and multivariate Cox proportional hazards regression analyses. Subsequently, we developed five machine-learning classifiers to predict 5-year survival. The discrimination of each model was assessed using the area under the receiver operating characteristic curve (AUC-ROC), and calibration was examined to ensure reliability. Kaplan–Meier analyses were performed to visualize survival distributions across key subgroups.
Results
This study included 1575 patients, of whom 372 and 1203 had VaM and VuM, respectively. The median patient age was 67 years, and the median tumor size was 2.4 cm. The 5-year survival rate of patients with VuM (45.4%) was significantly higher than that of patients with VaM (15.2%) (P < 0.001).
Conclusions
This study highlights the aggressive nature of rare GU melanomas and the importance of surgical intervention and caution in the use of chemotherapy and radiotherapy.
Keywords: Machine learning, melanoma, prognosis, survival analysis, vaginal neoplasms, vulvar neoplasms
PLAIN LANGUAGE SUMMARY
This study focused on rare and aggressive cancers known as VaM and VuM melanomas, which arise in the GU system. Using machine learning techniques, we analyzed patient data to identify factors affecting survival and created the first web-based tool to predict survival outcomes in these cancers. This study found that surgical treatment significantly improved survival rates, whereas chemotherapy and radiotherapy had mixed effects. Tumor size and marital status were identified as important predictors of survival. The developed web tool allows clinicians to input patient information and receive personalized survival predictions, helping guide treatment decisions. This study highlights the need for cautious use of therapies and offers innovative ways to personalize care for patients with these rare cancers of therapies and offers innovative ways to personalize care for patients with these rare cancers.
ARTICLE HIGHLIGHTS
This is the first study to develop a web-based predictive tool for vaginal and vulvar melanomas (VaM and VuM).
Machine learning algorithms were used to model the 5-year survival based on SEER data from 1,575 patients.
Surgical treatment significantly improved overall and cancer-specific survival in both VaM and VuM patients.
Radiotherapy was associated with poorer survival outcomes, highlighting the need for cautious use.
Larger tumor size was a negative prognostic factor for both cancers, and marital status also influenced outcomes.
The predictive tool enables individualized survival estimates and may assist in patient stratification and treatment planning.
1. Background
One out of every 41 (2.4%) women will develop a malignant melanoma at some point in their lives, ranking it as the sixth most prevalent cancer among women in the United States [1]. While melanoma typically manifests on the skin, it can also arise in noncutaneous areas such as the eyes and mucous membranes [2]. Mucosal melanoma can develop from mucous membranes in various regions of the body, including the head and neck, female genital organs, the anorectal area, and the urinary tract. Within mucosal melanomas, those affecting the GU system account for almost half (44.8%) of all cases [2]. This form of melanoma results from the migration of melanocytes to noncutaneous organs following an epithelial-mesenchymal transition of neural crest cells [3]. While cutaneous melanoma is typically linked to BRAF mutations, mucosal melanoma is associated with higher KIT or NRAS mutations or both [4]. Female GU tract melanomas are rare, accounting for less than 5% of all vaginal malignancies and between 0.2 and 0.8% of all melanomas and mostly arise from the vulva (76%) and vagina (19%) [5]. Chronic inflammatory diseases, viral infections, and chemical irritants have all been suggested as risk factors for GU melanoma in women [6].
Primary melanoma of the GU system is a rare, aggressive, and poorly understood disease that primary care physicians, urologists, or gynecologists may encounter [6]. Patients with Genitourinary melanoma commonly present with a mass, pain, bleeding, pruritus, ulceration, or urinary symptoms such as dysuria and altered urinary stream [3]. When compared with cutaneous melanomas and other types of gynecologic cancers, the clinical outcome for female GU melanoma is poor with a 5-year overall survival (OS) of 27% for VaM and between 8% and 58% for VuM [7,8]
Due to its rarity, there is limited data on prognostic factors and staging for VuM and VaM [9]. While vulvar melanomas were historically staged using the International Federation of Gynecology and Obstetrics (FIGO) system for vulvar squamous cell carcinoma, the American Joint Committee on Cancer (AJCC) tumor-node-metastasis (TNM) staging system for cutaneous melanomas has been found to be more accurate in predicting recurrence- and progression-free survival [8]. Similarly, the AJCC TNM staging system is considered superior for VaM compared to the FIGO staging system [9]. Additionally, the absence of prospective, randomized trials makes it challenging to establish comprehensive management guidelines [9]. Typically, patients are diagnosed late in the disease progression, leading to extremely poor prognoses [10].
Artificial intelligence (AI) includes ML, which focuses on computer data analysis and the refinement of learning algorithms. ML has successfully addressed complex problems, particularly in medicine, where it has been applied to medical image recognition, treatment support, and biomedical research [11–13].
This research sought to construct an individualized prognostic framework for vulvar and vaginal melanoma by integrating traditional survival analysis with advanced machine learning techniques. By identifying the clinical and pathological factors most strongly associated with long-term survival, we aimed to offer a tool that supports tailored treatment decision-making. The secondary goal was to translate our predictive models into a user-friendly, web-based application, thereby facilitating real-time risk estimation and fostering precision oncology in routine practice.
2. Methodology
We conducted our analysis using patient-level data from the National Cancer Institute’s SEER registry, which captures approximately 27.8% of the US population. We included individuals with a first primary diagnosis of vulvar or vaginal melanoma (ICD-O-3 code 8720/3) confirmed by histology and excluded cases without histological verification, those with prior malignancies, or missing essential data, resulting in 1,575 eligible patients. From the database, we extracted demographic variables (age at diagnosis, race, marital status, and year of diagnosis), tumor characteristics (lesion size dichotomized as <2 cm vs. ≥2 cm, and SEER summary stage categorized as localized, regional, or distant), treatment modalities (surgery, radiotherapy, and chemotherapy), and survival outcomes (overall survival (OS) and cancer-specific survival (CSS]). All statistical analyses were performed using R (v4.2.2) and the tidyverse, survival, survminer, and broom packages. Categorical comparisons were made using chi-square tests, and Kaplan–Meier methods with log-rank testing were used to evaluate OS and CSS differences. Independent prognostic factors were identified using multivariate Cox proportional hazards models, with a two-tailed P < 0.05 denoting significance.
To develop and validate predictive models for 5-year survival, we used Python’s scikit-learn library. Five algorithms were trained—Random Forest, Gradient Boosting, Logistic Regression, K-Nearest Neighbors, and a Multilayer Perceptron neural network—on standardized features (using StandardScaler) and an 80/20 train–test split. We quantified the impact of each predictor using permutation importance and then assessed discrimination using ROC curves and AUC metrics (roc_auc_score), alongside precision, recall, and F1-score from classification reports. Model robustness was further confirmed through 10-fold cross-validation and bootstrap resampling to generate mean performance estimates with 95% confidence intervals.
3. Results
3.1. Clinicopathological characteristics
The study population comprised 1575 patients. The median age was 67 years. A total of 40.4% of the patients had a tumor size < 2 cm, with a median tumor size of 2.4 cm. The largest racial group was white, comprising 88.6% of the cases. Furthermore, most cases were localized (57.8%). Most patients had surgery (89.4%) whereas only 7.7% underwent chemotherapy and 20.8% underwent radiotherapy. The demographic and clinical characteristics of the patients are shown in Table 1.
Table 1.
Clinicopathological characteristics.
| Category characteristics | Vagina (N = 372) | Vulva (N = 1203) | Overall (N = 1575) | P-value |
|---|---|---|---|---|
| Age | ||||
| Mean (SD) | 67.5 (12.5) | 65.0 (15.9) | 65.6 (15.2) | |
| Median [Min, Max] | 69.0 [28.0, 85.0] | 67.0 [14.0, 85.0] | 67.0 [14.0, 85.0] | 0.0049 |
| Race | ||||
| Asian or Pacific Islander | 31 (8.3%) | 51 (4.2%) | 82 (5.2%) | |
| Black | 39 (10.5%) | 59 (4.9%) | 98 (6.2%) | <0.001 |
| White | 302 (81.2%) | 1093 (90.9%) | 1395 (88.6%) | |
| Year of diagnosis | ||||
| 2000–2012 | 200 (53.8%) | 555 (46.1%) | 755 (47.9%) | 0.036 |
| 2013–2020 | 172 (46.2%) | 648 (53.9%) | 820 (52.1%) | |
| Marital status | ||||
| Married | 172 (46.2%) | 567 (47.1%) | 739 (46.9%) | 0.955 |
| Not married | 200 (53.8%) | 636 (52.9%) | 836 (53.1%) | |
| Stage | ||||
| Distant | 80 (21.5%) | 86 (7.1%) | 166 (10.5%) | |
| Localized | 162 (43.5%) | 748 (62.2%) | 910 (57.8%) | <0.001 |
| Regional | 130 (34.9%) | 369 (30.7%) | 499 (31.7%) | |
| Tumor size | ||||
| < 2 cm | 67 (18.0%) | 570 (47.4%) | 637 (40.4%) | <0.001 |
| > 2 cm | 305 (82.0%) | 633 (52.6%) | 938 (59.6%) | |
| Cancer directed surgery | ||||
| Not performed | 99 (26.6%) | 68 (5.7%) | 167 (10.6%) | <0.001 |
| Surgery performed | 273 (73.4%) | 1135 (94.3%) | 1408 (89.4%) | |
| Chemotherapy | ||||
| No | 314 (84.4%) | 1140 (94.8%) | 1454 (92.3%) | <0.001 |
| Yes | 58 (15.6%) | 63 (5.2%) | 121 (7.7%) | |
| Radiation | ||||
| No | 171 (46.0%) | 1077 (89.5%) | 1248 (79.2%) | <0.001 |
| Yes | 201 (54.0%) | 126 (10.5%) | 327 (20.8%) |
3.2. Survival analysis
As illustrated in Figure 1, the results of the survival analysis showed significant differences in OS and CSS (P < 0.001) between vaginal and vulvar melanomas. In particular, the survival rates for VaM were 15.2% for OS and 17.1% for CSS, whereas for VuM, the rates were substantially higher, at 45.4% for OS and 53.2% for CSS. As shown in Figure 2, among the VaM patients, those who did not receive radiotherapy had an OS rate of 18% and a CSS rate of 18.6%. In contrast, those who received radiotherapy had lower OS (12.8%) and CSS (15.6%) rates. VuM patients who did not receive radiotherapy had an OS rate of 47% and a CSS rate of 55.4%. In contrast, Figure 3 illustrated that patients who underwent radiotherapy had an OS rate of 31.1% and a CSS rate of 33.6%.
Figure 1.
(A) Kaplan–Meier overall survival (OS) for treatment; (B) Kaplan–Meier cancer-specific survival (CSS) for anatomical site.
Figure 2.
(A) Kaplan–Meier overall survival (OS) for treatment; (B) Kaplan–Meier cancer-specific survival (CSS) for radiation.
Figure 3.
(A) Kaplan–Meier overall survival (OS) for treatment; (B) Kaplan–Meier cancer-specific survival (CSS) for radiation.
3.3. Prognostic factors
To determine the possible independent prognostic factors for OS and CSS in patients with VuM and VaM, we employed a univariate Cox regression model and subsequently selected only the significant variables for multivariate analysis. Large tumor size is a negative prognostic factor for both OS and CSS in patients with VaM. In contrast, chemotherapy is a favorable prognostic factor for both OS and CSS, whereas surgery is beneficial only for OS Supplementary Figure 1. Unmarried VuM is a negative prognostic factor for OS, while larger tumor size and chemotherapy are poor prognostic factors for both OS and CSS. However, surgery was identified as a positive prognostic factor for both the OS and CSS Supplementary Figure 2.
3.4. Model performances and interpretability
The detailed performance metrics for all the ML algorithms are summarized in Supplementary Table 1. ROC curves of all the MLMs are displayed in Figure 4. The features that contributed the most to survival prediction are displayed in Figure 5. Notably, the vaginal model achieved a higher AUC (0.82) than the vulvar model (0.62), likely due to the more consistent prognostic patterns in the VaM group. In contrast, greater heterogeneity in clinical features and treatments within the VuM group may have reduced the model performance. The factors contributing to vaginal melanomas include chemotherapy, tumor size, and radiation. Moreover, marital status, tumor size, and cancer-directed surgery were the main contributing prognostic elements for vulvar melanomas. We have deployed an interactive Shiny application powered by our Random Forest algorithm that ingests patient-specific inputs (e.g., age, tumor size, stage, and treatment modalities) and computes individualized survival predictions. Upon submission of these variables, the tool returns an estimated survival duration in months, along with confidence intervals derived from the model’s internal validation process. Designed for speed and accessibility, clinicians and investigators can access the platform at https://sakhrashwayyat.shinyapps.io/Melanoma_GU/ to support prognostic assessments and treatment planning.
Figure 4.
( A) Receiver operating characteristic (ROC) curves of all machine learning models (MLMs); (B) permutation features importance (random forest classifier).
Figure 5.
(A) Receiver operating characteristic (ROC) curves of all machine learning models (MLMs); (B) permutation features importance (random forest classifier).
4. Discussion
4.1. Overview and clinical context
GU malignant melanoma is an uncommon yet aggressive condition with unfavorable outcomes [14]. The clinical presentation, diagnosis, and staging of this pathology vary tremendously among different patient groups. Due to their rarity and heterogeneity, staging and treating GU melanomas can be quite challenging [15]. This study utilized ML to examine a plethora of prognostic factors contributing to VaM and VuM disease etiology. Our study provides the first personalized web-based tool to predict the survival rates for patients with these rare types of cancer.
Regarding tumor size, VuMs typically manifest as smaller lesions, with prognosis primarily associated with depth rather than diameter [16]. A retrospective analysis of 100 cases of primary VuM revealed that increased tumor thickness was correlated with both reduced OS and lower disease-specific survival [17]. A study analyzing 33 patients with VuM confirmed Breslow depth as a strong predictor of recurrence and poor disease-free survival (DFS) [16]. However, our study categorized tumors based on size rather than depth, classifying them as <2 and >2 cm. Our findings indicate that larger tumors are indicative of poorer prognosis.
4.2. Efficacy of surgery
The management of GU melanoma is quite challenging owing to the limited number of cases encountered [18,19]. Nonetheless, the main treatment for GU melanoma is surgery [18,20]. The type of surgery, whether radical or conservative, does not influence patient survival [21]. One study classified primary tumor surgery into two categories: radical and wide excision. In this study, 26 patients underwent radical surgery, which could involve procedures such as radical vulvectomy, hemivulvectomy, hysterectomy, and vaginectomy, with or without oophorectomy or a combination of vulvectomy and partial vaginectomy. Meanwhile, 49 patients underwent wide local excision, which included procedures such as wide local excision with primary closure, flap closure, or graft closure [9]. A study conducted at the University of Texas reviewed the records of 37 patients with vaginal melanoma and found that those who underwent surgical procedures (such as pelvic exenteration or wide excision) had longer survival rates than those treated non-surgically with primary radiation therapy, chemotherapy, or both [21]. Consistent with these findings, our study demonstrated that vaginal and vulvar melanoma patients who underwent cancer-directed surgery showed significantly increased OS and CSS rates.
4.3. Controversial and emerging systemic therapies
The literature suggests that radiotherapy (RT) for advanced gynecological melanomas can reduce local recurrence rates and enhance survival [21,22]. In VuM, RT has been used as a preoperative neoadjuvant therapy to shrink tumors, facilitating more conservative surgeries [10]. However, RT for vulvar tumors has been technically challenging due to difficulties in targeting the external RT beam to the vulva without causing splatter and the high sensitivity of vulvar skin and surrounding mucosal tissue to RT effects. Notably, one study found no improvement in OS or recurrence-free survival in patients with VuM who received adjuvant RT [23]. The ongoing debate regarding the effectiveness of RT might be due to the rarity and limited number of cases. In our study, radiation was significantly associated with poor prognosis.
Furthermore, chemotherapy appears to be useful in some studies [23,24]. In a study involving 189 patients with mucosal melanoma, 63 received temozolomide plus cisplatin as a systemic adjuvant therapy for resected mucosal melanoma, which led to a significant increase in the OS rate compared with patients who only underwent surgery without chemotherapy or those treated with high-dose IFN-α2b (HDI) [24]. In our study, chemotherapy was the most important factor affecting the survival outcomes of patients with VaM and VuM. Our study found a significant association between chemotherapy and poor prognosis, which may be attributed to the considerable side effects experienced during treatment.
Given the poor prognosis of chemotherapy, it is crucial to consider the side effects of chemotherapy that patients may experience during treatment. One study has highlighted that patients can develop chemotherapy-induced peripheral neuropathy (CIPN), which significantly decreases their quality of life [25]. Dermatological side effects of chemotherapy are frequent in the treatment of women’s cancers and significantly affect quality of life, as measured by health-related quality of life (HRQL) scores. Patients should be counseled about these side effects before starting adjuvant or palliative chemotherapy. Dermatological side effects include palmoplantar erythrodysesthesia (PPE), keratosis follicularis, facial flushing, psoriasis-like plaques, urticaria, rhagades, petechiae, and cutaneous mycosis. Chemotherapy-induced hair loss (CIA) is a distressing side effect that is common in certain oncology treatment regimens. Although temporary, CIA can cause severe emotional distress and affect patient compliance, psychosocial stability, and overall patient satisfaction. Therefore, education, psychological support, physician interaction, and subspecialty referrals are highly recommended [25].
Immunotherapy, particularly techniques targeting T cell-mediated immunity, has shown promising results in treating patients with advanced-stage metastatic melanoma [26,27]. Therapeutic approaches such as IFN-γ, TNF-a, PD-1 therapy, and CTLA4 agents have demonstrated notable efficacy in this regard [26,28,29]. A recent review of cancer immunity has shown that although anti-CTLA4 and PD-L1/PD-1 agents have shown clinical success, it is vital to recognize that only a subset of patients exhibit durable responses to these treatments [30].
To the best of our knowledge, this study is the first to develop a web-based tool for predicting the outcomes of VaM and VuM. Utilization of a large national database ensures robust and widely applicable results. We employed various ML algorithms to accurately identify survival predictors, considering numerous patient and disease characteristics, for a comprehensive analysis. However, the retrospective design of our study limited the causal conclusions. Additionally, our analysis did not thoroughly account for treatment dosage variations, such as chemotherapy, radiotherapy, or other health conditions, which could significantly impact patient outcomes. The SEER database also lacks detailed information on recurrence, molecular profiles and does not include data on tumor thickness (e.g., Breslow depth), a known prognostic factor in melanoma, which limited our ability to assess this parameter directly. Moreover, all models were trained and validated using the SEER data. We recommend future external validation using independent datasets or prospective multiinstitutional cohorts to confirm the generalizability and clinical utility of the developed models.
5. Conclusion
Our study presents significant advancements in the management and understanding of VaM and VuM through the development of our first web-based predictive tool. This tool represents a novel approach to oncological prognosis, offering personalized survival predictions and supporting clinicians in making informed treatment decisions. These findings underscore the importance of surgical interventions and raise caution regarding the use of chemotherapy and radiotherapy, which are associated with varied impacts on survival. Chemotherapy, while beneficial in some contexts, has a mixed influence on prognosis, emphasizing the need for careful patient selection and management. Radiotherapy appears to be less beneficial, suggesting a potential reevaluation of its role in treatment protocols.
Supplementary Material
Author contributions
Sakhr Alshwayyat: Conceptualization, Formal analysis, Methodology, Software, Visualization, Writing—original draft. Zena Haddadin: Writing—original draft. Sara Haddadin: Methodology. Mustafa Alshwayyat: Writing—original draft. Tala Abdulsalam Alshwayyat: Formal analysis, Software. Muna Talafha: Writing—original draft. Hamdah Hanifa: Writing—original draft. Jihan Muhaidat: Supervision, Writing—review & editing.
Ethics approval and consent to participate
The ethical approval and informed consent statements are not applicable for this type of research. Authorization and data were obtained through the SEER website and database, respectively.
Disclosure statement
The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
References
Papers of special note have been highlighted as either of interest (•) or of considerable interest (••) to readers.
References
- 1.Wohlmuth C, Wohlmuth-Wieser I.. Vulvar melanoma: molecular characteristics, diagnosis, surgical management, and medical treatment. Am J Clin Dermatol. 2021;22(5):639–651. doi: 10.1007/s40257-021-00614-7 * A comprehensive modern synthesis of vulvar melanoma management, integrating molecular data. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Vyas R, Thompson CL, Zargar H, et al. Epidemiology of genitourinary melanoma in the United States: 1992 through 2012. J Am Acad Dermatol. 2016;75(1):144–150. doi: 10.1016/j.jaad.2015.10.015 [DOI] [PubMed] [Google Scholar]
- 3.Tacastacas JD, Bray J, Cohen YK, et al. Update on primary mucosal melanoma. J Am Acad Dermatol. 2014;71(2):366–375. doi: 10.1016/j.jaad.2014.03.031 [DOI] [PubMed] [Google Scholar]
- 4.van Engen-van Grunsven ACH, Küsters-Vandevelde HVN, De Hullu J, et al. NRAS mutations are more prevalent than KIT mutations in melanoma of the female urogenital tract—a study of 24 cases from the Netherlands. Gynecol Oncol. 2014;134(1):10–14. doi: 10.1016/j.ygyno.2014.04.056 [DOI] [PubMed] [Google Scholar]
- 5.Xia L, Han D, Yang W, et al. Primary malignant melanoma of the vagina: a retrospective clinicopathologic study of 44 cases. Int J Gynecol Cancer. 2014;24(1):149–155. doi: 10.1097/IGC.0000000000000013 [DOI] [PubMed] [Google Scholar]
- 6.Sanchez A, Rodríguez D, Allard CB, et al. Primary genitourinary melanoma: epidemiology and disease-specific survival in a large population-based cohort. Urol Oncol. 2016;34(4):166.e7–166.e14. doi: 10.1016/j.urolonc.2015.11.009 [DOI] [PubMed] [Google Scholar]
- 7.Hou JY, Baptiste C, Hombalegowda RB, et al. Vulvar and vaginal melanoma: a unique subclass of mucosal melanoma based on a comprehensive molecular analysis of 51 cases compared with 2253 cases of nongynecologic melanoma. Cancer. 2017;123(8):1333–1344. doi: 10.1002/cncr.30473 * Pivotal study distinguishing mucosal melanoma’s molecular profile from other melanoma types. [DOI] [PubMed] [Google Scholar]
- 8.Moxley KM, Fader AN, Rose PG, et al. Malignant melanoma of the vulva: an extension of cutaneous melanoma? Gynecol Oncol. 2011;122(3):612–617. doi: 10.1016/j.ygyno.2011.04.007 [DOI] [PubMed] [Google Scholar]
- 9.Seifried S, Haydu LE, Quinn MJ, et al. Melanoma of the vulva and vagina: principles of staging and their relevance to management based on a clinicopathologic analysis of 85 cases. Ann Surg Oncol. 2015;22(6):1959–1966. doi: 10.1245/s10434-014-4215-3 [DOI] [PubMed] [Google Scholar]
- 10.Rambhia PH, Scott JF, Vyas R, et al. Genitourinary melanoma. In: JF Scott, MR Gerstenblith, editor. Noncutaneous Melanoma. Brisbane (AU): Codon Publications; 2018. [PubMed] [Google Scholar]
- 11.Deo RC. Machine learning in medicine. Circulation. 2015;132(20):1920–1930. doi: 10.1161/CIRCULATIONAHA.115.001593 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Komura D, Ishikawa S.. Machine learning approaches for pathologic diagnosis. Virchows Arch. 2019;475(2):131–138. doi: 10.1007/s00428-019-02594-w [DOI] [PubMed] [Google Scholar]
- 13.Handelman GS, Kok HK, Chandra RV, et al. eDoctor: machine learning and the future of medicine. J Intern Med. 2018;284(6):603–619. doi: 10.1111/joim.12822 [DOI] [PubMed] [Google Scholar]
- 14.Khayyat A, Esmaeil Pour MA, Mousavi S, et al. Primary malignant melanoma of the genitourinary system: a systemic review and report of eight cases. Cureus. 2022;14(11):e30444. doi: 10.7759/cureus.30444[Mismatch] ** Provides a recent systemic review and new cases to expand insight on rare GU melanomas. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.DePalo DK, Elleson KM, Carr MJ, et al. Genitourinary melanoma: an overview for the clinician. Asian J Urol. 2022;9(4):407–422. doi: 10.1016/j.ajur.2022.01.003 ** Offers clinical-focused, up-to-date guidance on diagnosis and management of GU melanomas. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Heinzelmann-Schwarz AV, Nixdorf S, Valadan M, et al. A clinicopathological review of 33 patients with vulvar melanoma identifies c-KIT as a prognostic marker. Int J Mol Med. 2014;33(4):784–794. doi: 10.3892/ijmm.2014.1659 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Nagarajan P, Curry JL, Ning J, et al. Tumor thickness and mitotic rate robustly predict melanoma-specific survival in patients with primary vulvar melanoma: a retrospective review of 100 cases. Clin Cancer Res. 2017;23(8):2093–2104. doi: 10.1158/1078-0432.CCR-16-2126 * Large study identifying strong prognostic markers specific to vulvar melanoma. [DOI] [PubMed] [Google Scholar]
- 18.Dobrică EC, Vâjâitu C, Condrat CE, et al. Vulvar and vaginal melanomas—the darker shades of gynecological cancers. Biomedicines. 2021;9(7):758. doi: 10.3390/biomedicines9070758 * Recent review emphasizing diagnostic challenges and evolving treatment strategies. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Cuccia F, D’Alessandro S, Blasi L, et al. The role of radiotherapy in the management of vaginal melanoma: a literature review with a focus on the potential synergistic role of immunotherapy. JPM. 2023;13(7):1142. doi: 10.3390/jpm13071142 ** A recent and detailed review emphasizing emerging treatment combinations, particularly radiotherapy and immunotherapy. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Skovsted S, Nielsen K, Blaakær J.. Melanomas of the vulva and vagina. Dan Med J. 2017;64(2):A5346. [PubMed] [Google Scholar]
- 21.Frumovitz M, Etchepareborda M, Sun CC, et al. Primary malignant melanoma of the vagina. Obstet Gynecol. 2010;116(6):1358–1365. doi: 10.1097/AOG.0b013e3181fb8045 [DOI] [PubMed] [Google Scholar]
- 22.Wang D, Xu T, Zhu H, et al. Primary malignant melanomas of the female lower genital tract: clinicopathological characteristics and management. Am J Cancer Res. 2020;10(12):4017–4037. [PMC free article] [PubMed] [Google Scholar]
- 23.Janco JMT, Markovic SN, Weaver AL, et al. Vulvar and vaginal melanoma: case series and review of current management options including neoadjuvant chemotherapy. Gynecol Oncol. 2013;129(3):533–537. doi: 10.1016/j.ygyno.2013.02.028 [DOI] [PubMed] [Google Scholar]
- 24.Lian B, Si L, Cui C, et al. Phase II randomized trial comparing high-dose IFN-α2b with temozolomide plus cisplatin as systemic adjuvant therapy for resected mucosal melanoma. Clin Cancer Res. 2013;19(16):4488–4498. doi: 10.1158/1078-0432.CCR-13-0739[Mismatch] * One of the few RCTs addressing systemic adjuvant therapy for mucosal melanoma. [DOI] [PubMed] [Google Scholar]
- 25.Hsu HC, Tsai SY, Wu SL, et al. Longitudinal perceptions of the side effects of chemotherapy in patients with gynecological cancer. Support Care Cancer. 2017;25(11):3457–3464. doi: 10.1007/s00520-017-3768-7 [DOI] [PubMed] [Google Scholar]
- 26.Schreiber RD, Old LJ, Smyth MJ.. Cancer immunoediting: integrating immunity’s roles in cancer suppression and promotion. Science. 2011;331(6024):1565–1570. doi: 10.1126/science.1203486 [DOI] [PubMed] [Google Scholar]
- 27.Hodi FS, O’Day SJ, McDermott DF, et al. Improved survival with ipilimumab in patients with metastatic melanoma. N Engl J Med. 2010;363(8):711–723. doi: 10.1056/NEJMoa1003466 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Topalian SL, Hodi FS, Brahmer JR, et al. Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N Engl J Med. 2012;366(26):2443–2454. doi: 10.1056/NEJMoa1200690 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Wolchok JD, Kluger H, Callahan MK, et al. Nivolumab plus ipilimumab in advanced melanoma. N Engl J Med. 2013;369(2):122–133. doi: 10.1056/NEJMoa1302369 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Chen DS, Mellman I.. Elements of cancer immunity and the cancer-immune set point. Nature. 2017;541(7637):321–330. doi: 10.1038/nature21349 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.