**4. Discussion**

In this study, we evaluated the significance of metabolic 18F-FDG PET/CT volumetric parameters in predicting the overall survival of patients presenting for restaging. Our patient cohort had a recurrence in the primary site of disease, as well as distant disease. Regarding the multivariate and univariate parameters, patients that demonstrated higher MTV and TLG were at a higher risk of death. In relation to survival, only whole-body TLG was significantly associated. Our data did not demonstrate a positive correlation with other variables, such as age or the site of the primary tumor, and survival.

Metabolic PET parameters have been discussed as a potential benefit in the prognosis of solid tumors. This has been seen in head and neck cancers, lung cancer and gynecological malignancies [16]. There are several metabolic parameters measured in 18F-FDG PET/CT. A high metabolic volume in 18F-FDG PET scans is associated with poor prognosis and changes in metabolic activity can be used to monitor response to therapy [17]. Metabolic PET parameters have been demonstrated to be independent prognostic factors for progression-

free survival (PFS) and overall survival (OS) [18]. Metabolic PET parameters that are tumor-based, namely, MTV and TLG, have been shown to better represent the entire tumor and are closely associated with prognosis in various malignancies [18]. MTV and TLG have also been demonstrated to have a greater association with PFS and OS when compared to SUVmax. Metabolic parameters of tumor volumes have been demonstrated to be associated with prognostic values of overall survival and progression-free survival in esophageal cancer [19], head and neck cancer [20] and small cell lung cancer [21].

A study by Son et al. looked at the prognostic relevance of MTV and TLG in patients with malignant melanoma and found that pre-treatment MTV and TLG may be useful in risk-stratifying patients for likelihood of death and recurrence, with TLG being the best predictor [9]. Our findings were similar, where those patients with higher MTV, TLG and whole-body TLG were associated with overall survival. The difference between our study and theirs is that they looked at patients who presented for initial staging and our cohort of patients presented for restaging. Our study did not focus on progression-free survival, as a large proportion of our patients already had stage IV disease. The findings did however agree that metabolic parameters played a role in prognosis.

Another study by Ito et al., which looked at tumor volumes, demonstrated that wholebody MTV is a strong independent prognostic factor in determining which melanoma patients will respond to immunotherapy [8]. Our study did not look at the response to therapy but demonstrated that a higher whole-body MTV was associated with poorer survival. This agrees with another study that looked at treatment response and reported that those with more tumor involvement faired more poorly than those with less tumor involvement. This demonstrates that tumor volumes can be used to see which patients are more likely to benefit from intervention.

More recently, a study by Reinert et al. demonstrated the prognostic value of metabolic parameters in patients with melanoma regarding progression-free and overall survival [22]. This study demonstrated a positive correlation between MTV and TLG with overall and progression-free survival [22]. These findings were almost similar to our findings; however, this was done on a European population that presumably has better access to healthcare. In our population, patients typically present late and with more advanced disease [2]; therefore, in our population, having an additional risk stratification tool may guide clinicians regarding which therapies to use, more frequent follow-ups or earlier palliative treatment if necessary. The study also looked at other parameters, such as serological markers, lactate dehydrogenase and C-reactive protein. These specific parameters were not commonly reviewed in our patient cohort. Prognostication using 18F-FDG PET/CT was also reviewed by Schweigoffer-Zwink et al., who demonstrated that metabolic parameters in patients with advanced cutaneous melanoma were predictive for survival in melanoma patients undergoing immunotherapy [23]. This study also found that tumor-to-background values had a stronger predictive value than MTV and TLG. Our study did not look at tumor-to-background ratios but similarly demonstrated that metabolic PET parameters can be predictive in a resource-constrained setting such as ours.

As our study specifically looked at restaging, our findings were somewhat similar to the Albano et al. group in Italy, which reviewed patients with 18F-FDG PET/CT after surgery with suspicion for recurrence or metastatic disease post-surgical intervention; they found that imaging a positive scan was associated with an increased risk of disease progression and a negative study demonstrated longer survival than a positive one [24]. Metabolic parameters were not reviewed in this study; however, findings agree that a positive PET has prognostic outcomes in survival [24]. 18F-FDG PET/CT also has the added advantage that it can detect melanoma recurrence in asymptomatic patients prior to clinical detection and this was demonstrated in a study done at our center [25].

Survival in melanoma is dependent on the stage at diagnosis. The different factors that encompass staging, namely, tumor size, nodal involvement and metastases, have been evaluated for melanoma-specific survival. Overall survival tends to be poorer depending on the stage at diagnosis [26]. This correlates with our data, as most of our patients presented

for restaging at a later stage and had poorer outcomes. Due to mostly delayed presentation in our patient population, our study evaluated patients that demonstrated disease in the primary tumor at restaging with distant metastases. Our patients had significantly higher tumor volumes compared to what has been described by other authors [11,22].

The strength of this study was in demonstrating the value of metabolic parameters 18F-FDG PET/CT in restaging patients with malignant melanoma for prognostic purposes. The role in recurrent melanoma is ye<sup>t</sup> to be defined in a prospective study and would be beneficial to guide clinicians on potential clinical outcomes of patients, especially in recurrent disease.

The limitations of this study are that it is retrospective with a limited sample size of patients that presented to our hospital for restaging. Our study sample was based on the hospital records, which were not intended for research; therefore, challenges with incomplete records were encountered. A large proportion of our patients were also lost to follow-up, which also influenced our results. Difficulties in record keeping have been described in a resource-constrained setting similar to ours by Pirkle et al., who found that researchers in diverse settings struggle with record keeping. The authors mentioned that illegible notes or missing records can affect hospital care and research and mentioned the need for electronic records, which may assist with improving this [27]. Unfortunately, in our setting, medical records are still done on paper and only imaging is available electronically. A study done in a first-world setting, namely, in Taiwan, by Li et al. found that, although electronic medical records were available, most retrospective studies had a case number of fewer than 100 patients, with the average being 41 [28]. The lower case numbers were speculated to be due to the authors' preference of accessing paper records despite the availability of electronic records [28]. These studies demonstrate that lower case numbers in retrospective data are not unique to our population alone, but are seen in low-income and first-world countries with access to better record keeping [27,28]. Another limitation is the inability to correlate the histology of all the metastatic lesions to truly confirm melanoma metastases despite anatomical features of metastases. Our data demonstrated a positive correlation with tumor volumes and overall survival in this retrospective analysis of patients presenting for restaging. Prospective studies in patients with melanoma preventing for staging would be beneficial, as the bulk of known literature is retrospective.
