KEY POINTS
- This prospective study enrolled 53 newly diagnosed head and neck cancer patients; 47 completed baseline neurocognitive testing, 30 were assessed at six months and only 24 at 12 months. Mean age was 59.3 years, 87% were men, 45% had tongue tumors and 38% oropharyngeal tumors.
- A deep-learning deformable-registration framework automatically segmented the whole brain plus temporal, frontal, occipital and parietal lobes, hippocampus, cingulate and cerebellum. The model had been trained separately using 29 patients with co-registered CT and MRI.
- Incidental intracranial exposure was generally low but strongly region-dependent. Median mean dose was 0.60 Gy to the whole brain, 4.31 Gy to the cerebellum, 1.09 Gy to the temporal lobes, 1.08 Gy to the hippocampus, and only 0.28 Gy to the parietal lobes.
- Standardized cognitive scores did not deteriorate during the first year. Mean language scores changed from 0 at baseline to 0.26 at six months and 0.36 at 12 months; learning and memory increased to 0.51 and 0.50, while attention/processing speed/executive function reached 0.22 at both time points.
- No brain region showed a statistically significant association between mean dose and language, learning/memory or attention/processing/executive performance at either six or 12 months. Most Pearson correlation estimates were approximately −0.1 to +0.1.
- Age-adjusted analyses were similarly negative. The authors therefore found no evidence that the observed low-dose incidental brain exposure caused measurable neurocognitive deterioration during early survivorship.
- Interpretation is limited by substantial attrition, low statistical power at 12 months, heterogeneous head and neck subsites, frequent chemotherapy (83%), and the lack of statistical adjustment for repeat-testing practice effects. Late radiation-related cognitive effects may also emerge beyond the study's one-year observation period.
CLINICAL TAKEAWAY
These data are reassuring regarding the relatively low incidental brain doses typically received during many head and neck treatments, but they should not be converted into brain dose constraints or interpreted as proof of long-term safety. Larger cohorts with longer follow-up are needed, especially for superior head and neck targets where temporal-lobe and cerebellar exposure can be substantially higher.