The present retrospective investigation demonstrated that the recently constructed CARWL score, which combines CAR and SWL, is a reliable biological indicator that categorizes LA-NPC patients into discrete groups with considerably distinct rates of RIT occurrence after C-CRT. Specifically, those with a CARWL-2 score (CAR > 3.0 and WL > 5.0%) had a significantly increased risk (44.2%) of developing RIT compared to those with CARWL-1 (23.2%) and CARWL-0 (8.7%) scores (P < 0.001).

In addition to the novel parameter CARWL score, our multivariate analysis results, in line with the available literature, indicated that pre-C-CRT MMO < 41.3, T3-4 stage, N2-3 stage, concurrent chemotherapy cycles 2–3, mean MAD ≥ 34.3 Gy, and MAD 55.5 Gy > 40.0% were also significant independent predictive parameters for RIT. The pretreatment MMO, dose of MAD, T-stage, and N-stage are well-known risk factors and have been used for creating predictive nomograms for RIT [16, 17]. Variable dose constraints such as V42 Gy and mean 58 Gy for masseter muscle were defined as clinically relevant dosimetric cut-offs for RIT prediction as our result [6, 18]. Considering the influence of chemotherapy on RIT rates, a recently published meta-analysis indicated that chemoradiotherapy significantly (P = 0.0003) increased the prevalence of trismus (OR = 2.55, 95% CI = 1.53–4.23) compared to RT [11]. Our findings in this situation revealed that undergoing 2–3 cycles of concurrent chemotherapy, as opposed to just 1 cycle, was associated with a 2.06-fold higher likelihood of developing RIT (P = 0.011), which supports the meta-analysis results.

The most striking discovery of our study was the presentation of the recently 'introduced CARWL score's unique capacity to stratify LA-NPC patients into three groups with significantly distinct RIT risks (P < 0.001). Specifically, the 44.2% RIT rate experienced by the CARWL-2 patients was nearly two- and fivefold higher than the 23.2% of CARWL-1 and 8.7% of CARWL-0 patients. Motivated by the fact that hyperinflammation, as evidenced by a high CAR value, and the presence of SWL are frequently associated with higher rates of treatment failures and treatment-related toxicities [19, 20], we hypothesized that CARWL scores, unique combinations of these two parameters, could also effectively predict and categorize RIT rates in LA-NPC patients undergoing C-CRT. This hypothesis served as the foundation for our current study. The CARWL score was first presented as a prognostic indicator in stage IIIC NSCLC patients who underwent definitive C-CRT, as reported by Topkan et al. This research found that patients with stage IIIC NSCLC were stratified into three groups based on their CARWL score, which was used to assess OS and PFS. The OS and PFS were considerably shorter in patients with a CRAWL-2 score than those with CARWL-0 and CARWL-1 scores. The OS was 37.3 months for CARWL-0, 23.6 months for CARWL-1, and 12.8 months for CARWL-2 (P < 0.001). Similarly, the PFS was 14.2 months for CARWL-0, 11.4 months for CARWL-1, and 7.5 months for CARWL-2 (P < 0.001). The CARWL-2 score groups' respective 7.5 months PFS, and 12.8 months OS rates were remarkably close to those reported for metastatic patients in the NSCLC literature, demonstrating the robust prognostic potency of the CARWL scores. However, our research is the first to assess and show the usefulness of CARWL scores in determining the distinctive RIT risk levels of LA-NPC patients who undergo exclusive C-CRT treatment.

Explaining the correlation between a higher CARWL score and an increased risk of RIT is challenging due to the absence of similar study results. However, it may be logical to analyze the two elements of CARWL, namely the CAR and SWL, in the context of radiation-induced toxicities in HNC patients to offer sensible reasons for this link. CAR, the first component of CARWL, consists of CRP and albumin. CRP is a resilient biomarker of acute-phase response that signifies varying degrees of systemic inflammation, and its concentrations increase promptly after any inflammatory stimulation. Hypoalbuminemia is a constant companion of elevated CRP levels [21]. The primary reason for the inverse association between CRP and albumin levels is the inhibition of albumin production in hepatocytes brought on by CRP and its byproducts. Due to this robust inverse relationship, CAR levels are significantly elevated in response to even a single inflammatory stimulus. Hyperinflammation is a well-established state associated with decreased treatment tolerance, survival rates, and higher incidences of treatment-related toxicities [22]. While no research has specifically examined the effects of elevated CAR levels on the incidence rates of RIT, a recent study by Li and colleagues found that higher levels of CRP and lower levels of albumin were both independently linked to higher rates of nasopharyngeal necrosis after radical RT or C-CRT [23]. Given the similarities between the underlying pathophysiological mechanisms of nasopharyngeal necrosis and RIT, together with the evidence that heightened levels of CRP and decreased levels of albumin imply higher CAR levels, it is reasonable to hypothesize that a higher CAR level is linked to increased RIT rates.

Irrespective of the exact causal association between a higher CARWL score and increased RIT rates, our results led to categorizing LA-NPC patients into three discrete RIT risk groups: CARWL-0, CARWL-1, and CARWL-2. We contend that this risk stratification may be crucial to ensure the timely implementation of preventive measures in the CARWL-1 and CARWL-2 cohorts, including minimizing doses to the masticatory apparatus to the lowest feasible levels without compromising tumor control rates, as well as incorporating early onset exercise programs and instrumentation during and after CCRT. Additionally, we recommend stricter follow-up protocols for these patient groups, including shorter follow-up intervals and phone-call follow-ups. These contentions are in line with the findings of the meta-analysis of 11 randomized controlled trials reported by Wang and colleagues, which indicates that this approach stands out as the most effective method for preventing trismus [24].

The second component of the CARWL score is SWL, which signifies a weight loss of > 5% and serves as a marker for cancer cachexia.30 SWL occurring before or during RT or C-CRT implies malnutrition, which may compromise the capacity of damaged tissues to heal radiation-induced injuries, including the masticatory apparatus [25]. Moreover, acute radiation toxicities and SWL may aggravate each other, creating a detrimental vicious circle [20]. Patients with pre-cachectic or cachectic states also present with elevated CRP and decreased albumin levels [26].The SWL, reduced albumin, and elevated CRP are closely interconnected, primarily due to CRP's proinflammatory byproducts, including tumor necrosis factor-alpha (TNF-α), IL-1, IL-6, and interferon-γ (IFNγ) [22]. For example, any increase in CRP levels abruptly leads to stimulation of TNF and IL-6 levels which inhibits albumin synthesis. Therefore, it seems that the novel CARWL score integrates many proinflammatory, inflammatory, immune, and nutritional factors either directly (SWL, CRP, and albumin) or indirectly (TNF-α IL-1). Considering the role of these factors and the CRP byproducts in RIT pathogenesis, it is rational to anticipate that patients presenting with higher CARWL scores would experience higher RIT rates. Nevertheless, this comment should be acknowledged as a plausible scientific inference and supported by future research.

The current study has several limitations. First, this study is a retrospective and single-center cohort analysis, making it susceptible to selection biases. Second, these findings are specific to patients with LA-NPC who underwent definitive C-CRT and may not be generalizable to patients who initially received induction chemotherapy, those treated with RT alone for earlier-stage NPCs, or LA-NPC patients who cannot tolerate it. Third, the absence of crucial immune-inflammation and cachexia biomarkers, such as platelets, neutrophils, monocytes, lymphocytes, TNF-α, IL-1, and IL-6 in our study, limits our ability to establish a direct cause-and-effect relationship between a higher CARWL score and increased RIT incidence rates. Fourth, we only used single-point or interval measurements to determine the CAR and SWL cut-off values. However, CRP, albumin, and SWL levels may change markedly during and after C-CRT because of other radiation-related side effects, changes in the tumor burden, and the host's inflammatory status. Therefore, further studies may determine better CAR and SWL cut-offs at either of these time points, which better discriminate RIT rates across the three CARWL score groups. Thus, unless comprehensive and well-planned research investigating these matters supports the results reported here, it is advisable to regard them as hypothetical rather than definitive suggestions for therapeutic use.