**4. Discussion**

Recently, the prognostic value of nutritional status and inflammation has been de-bated in cancer patients [30,31]. Many studies have demonstrated that an impaired nutritional status and an increased inflammatory response are related to a worse prognosis in patients with different types of human cancer [32,33]. Different biochemical parameters have been proposed to evaluate the immune-nutritional status such as blood neutrophil, lymphocyte, monocyte, platelet count, neutrophil–lymphocyte ratio (NLR), lymphocyte–monocyte ratio (LMR), and platelet–lymphocyte ratio (PLR) [34–36]. Immune-nutritional scores such as the Prognostic Nutritional Index (PNI) and the Controlling Nutritional Status (CONUT) have also recently been evaluated in human cancer [20–26,37]. The latter has recently been implemented as a tool of nutritional screening, emerging as an independent prognostic factor for OS in several types of solid tumors [38] as well as in hematological cancer [39,40].

To our knowledge, our study is the first to investigate the potential role of the baseline CONUT score as a prognostic factor for PFS and OS in patients with advanced thyroid cancer (differentiated, medullary, or poorly differentiated thyroid cancer) treated with TKIs. In recent years, TKIs have become a pivotal therapy in many types of cancer [2]. These drugs are scheduled for chronic administration and are often associated with systemic side effects favoring malnutrition (i.e., diarrhea, nausea, mucositis), if not properly managed [8].

Our cohort of patients showed an overall baseline good nutritional status, with a CONUT score suggestive of moderate malnutrition only in 5/42 (11.9%) patients, while none had severe grade of malnutrition (highest CONUT score recorded equal to 7). Nevertheless, a significant correlation between immuno-nutritional status and clinical outcome was observed. Using ROC curve analysis, we found that a cut-off of 3 was able to predict the response to treatment with 100% of specificity and 70% of sensitivity for PFS (AUC 0.980, *p* < 0.0001), and 93.7% of specificity and 68.8% of sensitivity for OS (AUC 0.840, *p* = 0.001). It is important to underline that the two groups of patients (CONUT score <3 and ≥ 3) were similar for clinical and pathological features that could have an impact on the clinical outcome. Nevertheless, the number of patients that took the full dosage did not differ between the two groups at the beginning of TKI treatment as well as at the last follow-up. Since TKI dosage is mostly related to the adverse events, we can assume that the tolerance to therapy was also similar in the two groups of patients.

The optimal cut-off of the CONUT score varied across different studies [38]; in our study, it was higher compared to that found in patients with small cell lung cancer [20], but lower compared to that calculated in other cancer studies [21–23,29]. This difference is probably due to the different nutritional status and pathogenic mechanisms underlying the various types of cancer.

Multiple potential mechanisms underlying the relationship between parameters included in the CONUT score and cancer prognosis have been widely analyzed. The serum albumin reflects both the nutritional and the inflammatory status and is considered as a prognostic factor in several cancers [10–12]. It has been postulated that pro-inflammatory cytokines such as interleukin 6 (IL-6) and tumor necrosis factor alpha (TNF-α), which modulate albumin hepatic synthesis, are associated with lower serum concentrations of this protein [41,42]. TNF-α induces gluconeogenesis and the lipid and muscle protein

catabolism. It also stimulates the production of reactive oxygen species (ROS) in tissues, leading to the activation of ubiquitin-proteasome pathway and inducing muscle protein catabolism [43]. Moreover, in preclinical studies, it has been demonstrated that TNF-α and interleukin 1 (IL-1) penetrate the hematoencephalic barrier, causing an anorexigenic effect [44]. The protein deficit causes a downregulation of hormonal and antibody synthesis, leading to a suppression in B-cell differentiation and T-cell activation [45].

There are many assumptions about the relationship between serum cholesterol and prognosis in cancer [46–48]. Okuyama et al. [49] and many other authors have demonstrated a correlation between hypocholesterolemia and worse prognosis in non-small cell lung cancer, gastric cancer, and localized renal cell carcinoma [50–52]. It seems that tumoral cells can determine a reduction in serum cholesterol through many mechanisms. such as an increased uptake into the cells caused by the exposition of low-density lipoprotein (LDL)-cholesterol receptors on the cell membrane [53].

The lymphocyte count is widely used in prognostic scores since the immune response against cancer largely depends on lymphocytes. CD4+ and CD8+ lymphocytes are implied in the prevention of neoplastic proliferation and invasion, CD8+ T cell counts are consistently associated with better survival in many types of cancer [54–56].

Interestingly, the CONUT score cut-off found in our study was the only independent parameter associated with PFS (HR 12.2, *p* < 0.0001) and OS (HR 23.55, *p* < 0.0001) by multivariate stepwise Cox-regression analysis, suggesting a possible role of nutritional status in the clinical outcome of thyroid cancer patients with advanced disease treated with anticancer therapy.

The study has a few limitations such as the presence of multiple histotypes of thyroid cancer, the use of different TKIs, the small sample size, and the retrospective design of the study.

However, the study has several strengths such as standardized management in the same institution with detailed information regarding diagnosis, treatment, and follow-up. Finally, to our knowledge, this is the first study to have evaluated the role of CONUT score as an immuno-nutritional tool in the clinical outcome of advanced thyroid carcinoma treated with TKIs.
