Improved detection of recurrent laryngeal tumor after radiotherapy using ¹⁸FDG-PET as initial method

Background and purpose: Timely detection of recurrent laryngeal tumor after radiation is an important predictive factor for curation as well as preservation of laryngeal function. Direct laryngoscopy under general anesthesia with taking of biopsies is the standard diagnostic procedure to detect recurrence when suspicion is raised. This, however, is an invasive and potentially damaging technique. Hence, a non-invasive diagnostic procedure, such as ¹⁸FDG-PET to stratify patients for direct laryngoscopy could be useful. ¹⁸FDG-PET is interpreted visually so that observer variation may affect clinical practice. In the present study, we therefore investigated this aspect of reproducibility. Patients and methods: Thirty consecutive patients suspected of recurrent laryngeal carcinoma after radiotherapy underwent ¹⁸FDG-PET and direct laryngoscopy under general anesthesia with taking of biopsies. ¹⁸FDG-PET scans were reported by nine nuclear medicine physicians and residents, using a three-point scaling system. The reference was the absence or appearance of a local recurrence in the 12 months following ¹⁸FDG-PET. Results: Eight patients had biopsy proven recurrent laryngeal carcinoma. Sensitivity of ¹⁸FDG-PET was 88% (95% CI 53-98%) and specificity was 82% (95% CI 62-93%). The observers had a moderate to reasonable agreement (weighted kappa 0.45 (95% CI 0.20-0.69)) vs. the clinical gold standard and interobserver kappa was 0.54 (95% CI 0.40-0.69). Conclusion: ¹⁸FDG-PET seems to be a promising technique to detect recurrent laryngeal carcinoma after radiotherapy, and selecting patients for direct laryngoscopy. This will avoid futile invasive procedures. Interobserver agreement and variability is reasonable using this technique, but training is necessary. Studies comparing different strategies to select patients for direct laryngoscopy in case of suspected recurrence are warranted.