Session Item

Optimal management of head and neck cancer of unknown primary (HNCUP) is still controversial. Radiotherapy (RT) is a main stem of HNCUP treatment. Also the choice of treatment volumes is controversial, in particular whether RT should be extended from the unilateral (UN) to the bilateral neck (BN) and whether or not include the pharyngeal mucosa (PM). A detailed understanding of treatment failure patterns is therefore essential.

This is a retrospective series of HNCUP patients (pts) treated between 2005 and 2020. Pts, tumor and treatment characteristics have been correlated to survival outcomes: overall (OS), progression free (PFS), local progression free (LPFS), regional recurrence free (RRFS) and metastases free (MFS) survival. LPFS has been defined as the time elapsed between end of treatment and appearance of the primary site of disease. MRI images of local/regional recurrences have been co-registered with RT images and plan and the area of recurrenc e(GTVrec) has been defined according to the prescribed dose received as “in-field” (> 80% of volume included in the 95% isodose), “out-of-field” (< 20% of volume included in the 95% isodose) or “marginal”(intermediate situation).

Of the 52 pts with HNCUP 7 (13.7%) had a local +/- regional recurrence. Nodal presentation was unilateral in 67.3% of pts and the majority of pts (78.8%) underwent a nodal dissection and adjuvant RT. RT volume was UN and BN in 28.8% and 71.2% respectively; in 64.4% of pts RT volumes included also the PM.  Mean dose to the high dose volume was 64.8 Gy, while mean dose to the low dose volume was 53.2 Gy. Concomitant chemotherapy was used in 53.8% of cases. None of the variables analysed correlated with OS. Presence of extranodal extension was the only variable associated with PFS [p<0.000 HR 23.2 (CI 4.9-108.8)] and MFS [p<0.004 HR 11.98 (CI 2.2-65.93)]. LPFS was correlated with a synchronous nodal progression [p<0.001 HR 13.01 (CI 2.9-59.3)] as no event of RRFS was seen in patients with no appearance of the primary site. The primary site was oropharynx, oral cavity and skin in 2, 4 and 1 pts respectively. Three pts had also a nodal recurrence. After image co-registration 40% of relapses were in-field, 10% marginal and 50% out-of-field. No differences were found in terms of RT volumes (UN versus BN, PM yes versus no) and evidence of relapse. Figure 1 summarizes the characteristics of patients with local/regional recurrence. Figure 2 shows an example of image co-registration.

Adequate staging procedures can lead to the identification of the primary site in most cases. Our data confirms  thats in cases of real HNCUP an intensified multimodal treatment and an extensive nodal/mucosal coverage with RT have shown no benefit on survival outcomes. ENE is closely linked to worse PFS and MFS and should prompt the investigation of new therapeutic approaches. Due to the lack of clinical prognostic factors we support molecular and radiomic research in better stratification of HNCUP.