Discussion

Nasopharyngeal carcinomas, particularly the undifferentiated type, which is the most common histological subtype, exhibit epidemiological, clinical and management characteristics that vary by region and population. Our study of 50 patients provides insights into these features in our Mediterranean context, showing an incidence consistent with intermediate-frequency regions and a demographic profile similar to published data. The male predominance with a sex ratio of 2:1 (66% men) aligns with literature reporting marked male predominance. ^{Reference Chang, Ye and Zeng14} The mean age at diagnosis was 46 years, within the typical range described, which includes two incidence peaks: an early peak between 15 and 25 years and a later peak around 50 years. ^{Reference Chang, Ye and Zeng14,Reference Busson, Ooka and Corbex15} In our cohort, age was not significantly associated with the development of hypothyroidism (p = 0·515). Cervical lymphadenopathy was the most common presenting symptom (80%), consistent with prior reports, while rhinological (40%) and otological (36%) symptoms often reflected locoregional tumour extension. Diagnosis was frequently delayed due to the deep tumour location. ^{Reference El Korbi, Ben Tkhayat and Bouatay16,Reference Saleh-Ebrahimi, Zwicker and Muenter17} Most patients presented with advanced TNM stages (III-IVB: 80%), in agreement with the literature. ^{Reference EAR, Irekeola and Yean18}

Regarding primary hypothyroidism after radiotherapy, it was observed in 66% of patients over a 60-month follow-up, defined by elevated TSH (>4·2 mUI/L) and low FT4 (<10 pmol/L). This incidence is higher than that reported in some studies, where rates generally range from 40% to 50%. ^{Reference Chyan, Chen and Shugard19,Reference Diaz, Jaboin and Morales-Paliza20} The cumulative incidence of hypothyroidism post-radiotherapy in our cohort was 6·1% at 12 months, 42·4% at 24 months and 90·9% at 36 months. These findings partly concur with literature data. Yoden et al. reported a 40% incidence at 24 months but less marked long-term increase, whereas Alterio et al. described a slower progression with 30% at 12 months and 50% at 36 months. ^{Reference Jereczek-Fossa, Alterio and Jassem12,Reference Prayongrat and Lertbutsayanukul21} The median time to hypothyroidism onset was 26 months (IQR 18–30), ranging from 6 to 42 months, comparable to Tell et al.’s median of 24 months, though some studies report peak incidence between 12 and 24 months post-irradiation. ^{Reference Mercado, Adelstein and Saxton22–Reference Murthy, Narang and Ghosh-Laskar24} This relatively long latency reflects the progressive radiation-induced damage mechanisms, including vascular and cellular injury leading to delayed thyroid dysfunction. ^{Reference Rai, Kumar and Mahobiya25}

The average thyroid gland volume in our study was 16·24 ± 6·74 cm³ (range 7–38·2 cm³), consistent with adult reference values between 15 and 18 cm³. ^{Reference Chaudhary and Bano26} Thyroid volume is an important factor, as smaller glands may be more susceptible to radiation damage at equivalent doses. ^{Reference Zhou, Chen and Shen27,Reference Nowicka, Tomasik and Papis-Ubych28} Analysis of predictive factors for post-therapeutic hypothyroidism showed female sex to be significant (p = 0·035), indicating women are more prone to develop primary hypothyroidism after treatment, consistent with Zhai et al.’s findings of female sex as an independent risk factor. ^{Reference Zhai, Lyu and Ni29} This increased susceptibility may relate to biological, hormonal and immunological sex differences, including a typically smaller thyroid volume in women, resulting in higher radiation dose per tissue unit, and direct effects of female sex hormones such as estrogens on thyrocyte function and oxidative stress response. ^{Reference Chaudhary and Bano26,Reference Lamichhane, Pant and Lamichhane30} Hypothyroidism occurrence was also significantly associated with advanced tumour stage (p = 0·007) and extensive nodal involvement (p = 0·003). The nodal extension (N stage) showed good predictive performance (p = 0·006; AUC = 0·74; 95% CI = 0·602–0·878), with sensitivity and specificity of 63·6% and 58·8%, respectively, for threshold ≥ N2. Overall, the TNM stage was a statistically significant predictor (p = 0·001) with an AUC of 0·784 (95% CI = 0·66–0·909), identifying stage ≥ IVa as the optimal cut-off (sensitivity 69·7%, specificity 76·5%). These results align with literature where higher nodal stages requiring larger irradiation volumes increase hypothyroidism risk due to greater thyroid radiation exposure. ^{Reference Prayongrat and Lertbutsayanukul21,Reference Zhai, Lyu and Ni29} The relationship between thyroid volume and post-radiotherapy hypothyroidism is well documented. Smaller thyroid volumes (<12·82 cm³) are linked to higher hypothyroidism rates (75% versus 37% for larger volumes). ^{Reference Zhou, Chen and Shen27,Reference Nowicka, Tomasik and Papis-Ubych28} Dosimetric parameters also correlate strongly with hypothyroidism risk. Vogelius et al. reported a mean thyroid dose of 45 Gy as a threshold associated with a 50% hypothyroidism risk, increasing beyond this dose. ^{Reference Vogelius, Bentzen and Maraldo10} Similarly, Bhandare et al. observed a 16% rise in incidence when mean dose exceeded 45 Gy. ^{Reference Bhandare, Kennedy and Malyapa31} In our cohort, patients developing hypothyroidism received significantly higher mean thyroid doses (59·8 ± 4·27 Gy) compared to those without hypothyroidism (55·19 ± 3·41 Gy; p < 0·001), reinforcing a dose-effect relationship. The mean thyroid dose predicted hypothyroidism well (p = 0·001) with an AUC of 0·782 (95% CI = 0·651–0·912), and an optimal cut-off at 56·67 Gy yielded 75·8% sensitivity and 64·7% specificity, consistent with thresholds reported in similar studies. ^{Reference Chyan, Chen and Shugard19}

Dose–volume parameters further highlighted the importance of volume irradiated at higher doses. The percentage of thyroid volume receiving 60 Gy (V60) showed excellent predictive performance (p < 0·001; AUC = 0·934; 95% CI = 0·87–0·999). A V60 threshold of ≥34·87% had 90·9% sensitivity and 82·4% specificity for hypothyroidism risk. This supports that larger thyroid volumes exposed to high doses substantially increase hypothyroidism risk, emphasising dose–volume constraints in treatment planning. ^{Reference Nowicka, Tomasik and Papis-Ubych28,Reference Bhandare, Kennedy and Malyapa31} Other volumes spared (VS40, VS45) had acceptable but lower predictive value (AUCs 0·774 and 0·793, respectively), with VS40 specificity at only 58·8%, limiting clinical utility. Notably, the thyroid volume spared from 60 Gy (VS60) was an excellent protective factor, with a threshold of 9·9 cm³ yielding 93·9% sensitivity and 100% specificity. This aligns with literature underscoring VS60 as a key parameter in preserving thyroid function post-radiotherapy and outperforming lower-dose spared volumes. ^{Reference Zhai, Lyu and Ni29,Reference Zhou, Chen and Tao32} Although radiotherapy remains the main risk factor for hypothyroidism, certain systemic treatments may exacerbate thyroid damage. Chemotherapy agents can increase radiation-induced inflammation or fibrosis of the thyroid, impairing hormone production. ^{Reference Yeung, Chiu and Vassilopoulou-Sellin33} In our study, neoadjuvant chemotherapy with 5FU-CDDP was significantly associated with higher hypothyroidism prevalence (42·4% vs. 5·9%; p = 0·008), supporting a contributory role of chemotherapy in thyroid dysfunction among nasopharyngeal carcinoma patients treated with radiotherapy. ^{Reference Zhou, Chen and Tao32,Reference Yeung, Chiu and Vassilopoulou-Sellin33}