Comparison of subjective evaluation versus objective algorithm in the interpretation of follow-up FDG-PET/CT scans after radiochemotherapy in head and neck cancer patients.

Aim Retrospective evaluation of serial FOG-PET/CT scans in head and neck squamous cell cancer (HNSCC) patient's follow-up after primary radiochemotherapy (RCTx), to assess the diagnostic accuracy of an experienced observer vs. an objective classification compared to standard clinical follow-up examinations. Methods Sixty-nine patients with locally advanced HNSCC were included, who received curative RCTx. Follow-up included serial FOG-PET/CT at the following time intervals t1: <= 270 d, t2: 271-540 d, t3:> 540 d after curative RCTx. The likelihood to detect local recurrences, nodal and distant metastases were compared between (i) experienced observer, (ii) an objective classification system by Zundel et al. [25], and (iii) routine clinical follow-up examinations. Results Twenty-two local recurrences, 7 nodal and 17 distant metastases were recorded during the follow-up. The diagnostic accuracy for local recurrence of the experienced observer vs. objective classification was 78 % vs. 77 % for t1, 83 % vs. 79 % for t2 and 100% vs. 84 % for t3. The classification (ii) and the conventional follow-up (iii) resulted in a relatively high amount of equivocal findings reducing the diagnostic accuracy. Conclusion Evaluation of FDG-PET/CT by an experienced observer in follow-up of HNSCC patients after curative RCTx resulted in the highest diagnostic accuracy in comparison to an objective classification and to routine clinical examination. HNSCC is a malignant tumor with a high likelihood of recurrence, especially in the first two years after curative RCTx. Early detection of recurrence is of high clinical importance, since there are several effective second line therapies that may have curative potential in some patients.