Uveal melanoma, from primary tumor to macrometastasis

Uveal melanoma (UM) is a rare but deadly intraocular malignancy. As the most commonly occurring primary intraocular cancer in adults, it poses significant challenges in prognostication and treatment due to its high propensity for developing late-stage metastasis. Moreover, since a 100% accuracy rate is virtually impossible, improvements can still be made to current clinical tools used for prognostication, cellular composition assessment, detection of dormant micrometastatic tumor cells, and evaluation of early treatment impact. Consequently, the focus of this thesis is to improve the prognostication, highlight important cellular features of both primary tumors and micrometastases, and shine a light on the importance of early treatment of the primary tumor.

In paper I, we elaborated on the diagnostic importance of uveal melanoma involvement of the ciliary body, monosomy 3, as well as the size, clinical and histological factors of the tumor. Furthermore, these factors were combined with the American Joint Committee on Cancer and the Cancer Genome Atlas models to develop a new prognostic classification system. Clinicopathological factors of two patient groups, consisting of 1796 patients, were compared between those with, and without, involvement of the ciliary body and monosomy 3. We found that involvement of the ciliary body, tumor diameter, and patient age at diagnosis to be independent predictors of monosomy 3. Moreover, tumor diameter, age at diagnosis, male sex, involvement of the ciliary body, and monosomy 3 independently predicted metastasis. Our proposed prognostic classification system, combining sex, patient age, extraocular extension, involvement of the ciliary body, monosomy 3, tumor size, and 8q (optional), outperformed both the American Joint Committee on Cancer 4 T-categories and the Cancer Genome Atlas groups A-D in validation cohorts.

In paper II, we digitally measured and examined the cellular composition of different melanocytes in tumors, nevi, and metastases from enucleated eyes and resected liver tissue retrieved from the archives of Ophthalmic Pathology section at St. Erik Eye Hospital. We identified 4 different variations of melanocytes in normal choroid, primary UMs and liver metastases, which would correspond with previously described normal choroidal melanocytes, and malignant melanocytes of the spindle A, spindle B, and epithelioid type. Contrasting most previous studies performed on tumor cell morphology and staining patterns, where manual assessments were performed, we used the more objective approach of computerized analysis. Previously shown superior to manual assessment regarding reproducibility and prognostic utility. We also analyzed the expression of BRCA1 associated protein-1 (BAP-1), Indoleamine 2,3- dioxygenase (IDO), Insulin-like growth factor-1receptor (IGF-1R), and T-cell immunoreceptor with Ig and ITIM domains (TIGIT) as well as examined the area of the nucleus, the nucleus perimeter, nucleus maximum and minimum caliper, nucleus eccentricity, and nucleus to cell area ratio. Consequently, we demonstrated that spindle A cells are more commonly located in the center of primary UM, whilst spindle B cells, together with epithelioid cells, are more common in the base of UM - an area recognized as the initial path for hematogenous spread of tumor cells. No normal melanocytes or spindle A cells were identified in liver metastases, where only spindle B or epithelioid cells were seen. In addition, we demonstrated diverse expressions of BAP-1, IDO, IGF-1R, and TIGIT between the different subsets of melanocytes, highlighting the fact that 4 different cell types, with distinct levels of protein expressions, can be identified during the progression of uveal melanoma.

In paper III, we examined the presence of dormant micrometastatic tumor cells in autopsy tissue from patients diagnosed with primary UM, without any clinical or radiological signs of macrometastatic disease. Data from over 4000 patients diagnosed at St. Erik Eye Hospital over a period of 60 years was examined together with data from autopsy registers. In the end, 11 subjects were identified as suitable candidates. Afterwards, tissue blocks obtained at autopsy were collected and examined with hematoxylin and eosin staining (HE) and immunohistochemistry (IHC). Thereafter, we analyzed of a subset of patients with immunomagnetic separation (IMS). Conclusively, we found micrometastases in several distant organs in 5 of 5 patients with coexisting macrometastases, as well as in 5 of 6 patients without. All micrometastases had very low, or no, proliferative activity when examined with the immunohistochemical marker M MIB1-IgG1 antibody against Ki67 (MIB1). With this finding, we suggest that micrometastases are present in practically every patient diagnosed with primary UM, and that macrometastases indeed do develop from micrometastases.

In paper IV, we wanted to determine the impacts of delayed treatment on overall survival. As uveal melanoma shows a high metastatic rate, potential effects of delayed treatment regarding prognosis has been under debate. Thus, we performed a retrospective cohort study on 1256 patients diagnosed with UM and collected data on their diagnosis and treatment dates. The patients were split into two different groups based on the interval between diagnosis and treatment, namely a delay in treatment of over, and under, one month. Prognostic effects of the interval between the two patient groups were evaluated. Consequently, we found that patients with a treatment delay of at least one month has both shorter overall survival and disease-specific survival in stages II and III. In addition, there was also an increased hazard ratio for metastatic death in both univariate and competing risk regression models. When examined with a Markov multi-state model, those with a treatment delay of over one month showed a hazard ratio of 1.45 (95% Confidence Interval (CI) 1.12-1.89) for evolution to metastatic death. Interestingly, when performing competing risk incidence analyses based on treatment timing, we were not able to prove any statistically significant disparities in the incidence of any cause of death other than metastases. We conclude that treatment delays of uveal melanoma are linked with reduced overall and disease-specific survival rates, leading to the suggestion that primary treatment should be initiated with the shortest possible delay after diagnosing UM.

List of scientific papers

I. VT Gill, S Sabazade, C Herrspiegel, et al. A prognostic classification system for uveal melanoma based on a combination of patient age and sex, the American Joint Committee on Cancer and the Cancer Genome Atlas models. Acta Ophthalmologica. 2023;101:34-48. https://doi.org/10.1111/aos.15210

II. Stålhammar G, VT Gill. Digital morphometry and cluster analysis identifies four types of melanocyte during uveal melanoma progression. Communications medicine. 2023;3(1):60. https://doi.org/10.1038/s43856-023-00291-z

III. VT Gill, E Norrman, S Sabazade, A Karim, E Lardner, G Stålhammar. Multiorgan Involvement of Dormant Uveal Melanoma Micrometastases in Postmortem Tissue From Patients Without Coexisting Macrometastases. American Journal of Clinical Pathology. 2023;160:164-174. https://doi.org/10.1093/ajcp/aqad029

IV. A Moghadam, VT Gill, S Sabazade, A Hagström, G Stålhammar. The Prognostic Significance of Treatment Delays on Uveal Melanoma Survival. [Manuscript]

All previously published papers were reproduced with permission from the publisher under a creative commons license.