Stereotactic radiotherapy in the management of oligometastatic and recurrent head and neck cancer: a single-center experience

For synchronous de novo oligometastatic patients, the metastatic site would be treated after treating the primary disease. Patients with extracranial metastases were treated with stereotactic body radiotherapy. Briefly, each patient received computerized tomography scan for planning purposes (P-CT) on a 16-slice CT scanner (Brilliance CT Big Bore Oncology, Philips Medical Systems, Inc., Cleveland, OH, USA) with vacuum cushions in breath-hold for thoracic or upper abdomen targets whenever possible. In cases where patients could not hold their breath, they underwent 4D P‑CT to delineate the moving target lesions with an additional abdominal compression being applied as passive motion management.

Patients with cranial metastases received stereotactic radiosurgery (SRS) or fractionated radiotherapy (SFR). Briefly, patients would receive histological confirmation, either by means of open surgery or stereotactic biopsy, followed by a P-CT for simulation in a stereotactic frame.

The radiation dose for SBRT/SRS/SFRT was prescribed to the isodose line (90–67%) to achieve dose-inhomogeneity in the center of the tumor [20]. The typical number of fractions would be 1–12, depending on the site, tumor size, and motion management. Lesions were counted separately if they had separate planning target volumes (PTV).

Thirty-one patients were identified in the local register with oligometastatic or recurrent head and neck tumors who received stereotactic radiotherapy throughout their treatment and met the inclusion criteria. The median prescribed dose was 45 Gy (range 60–24 Gy) with a median equivalent dose of 2 Gy per fraction (EQD2 _{α/β 10}) = 88 Gy (range 93.8–31.2 Gy) and median three fractions (range 1–12 fractions).

The treatment-related toxicities were collected from the documentation of the treating physician, and their severity was graded according to Common Terminology Criteria for Adverse Events (CTCAE) version 5.

The local ethics committee approved the current retrospective study (RWTH Aachen University, Faculty of Medicine, Ethics Committee “EK-22-440”). As an individual informed consent was obtained from each patient before the treatment, a separate consent for the study was not necessary.

The statistical analysis and graphics were executed using the R software version. The OS time was defined as the interval from the diagnosis of OMD to the time of death or censoring. Progression-free survival (PFS) was defined as the interval from the end of the radiation treatment until any site disease progression or censoring. Local control (LC) was defined as the interval from the end of the radiation treatment to the radiological progression of the irradiated lesion or censoring. Systemic therapy-free time (STFS) was defined as the interval from the end of radiation treatment to the introduction of systemic therapy by disease progression. The log-rank test was used for univariate analysis, and the Cox regression test was used for multivariate analysis. Time-dependent receiver operating characteristic curve (ROC) analysis was applied to calculate the cut-point with the most robust statistical significance.

Patients’ characteristics are provided in Table 1. In total, 48 lesions were treated, including 37 pulmonary metastases, 5 mediastinal lymph node metastases, 2 spine metastases, 1 regional recurrence, 1 brain metastasis, 1 adrenal gland metastasis, and 1 renal metastasis. Twenty-six patients had squamous cell cancer histology with 41 lesions. Five patients with a nonsquamous cell cancer histology had 7 lesions: 2 patients had adenoid cystic carcinoma, 2 patients had adenocarcinoma, and 1 patient had mucoepidermoid carcinoma. Of the 26 patients with a squamous cell cancer histology, 16 patients were p16-negative, 7 patients were p16-positive, and the p16-status for 3 patients was unknown.

Regarding the type of OMD, 28 patients had de novo OMD; this included 8 patients with synchronous OMD and 23 patients with metachronous OMD, including 1 patient with regional recurrence. Three patients previously had metachronous polymetastatic disease, received systemic therapy, converted to OMD (induced OMD), and subsequently received stereotactic radiotherapy to the oligometastatic lesions.

After 28.5 months of median follow-up, 12 of 31 patients had passed away, and the median overall survival time was 33 months. The OS rate at 1 year was 77.9%, and at 3 and 5 years it was 36.5% (Fig. 1a). The median OS for de novo OMD and induced OMD were 33 and 24.9 months, respectively.

Fifteen patients experienced a systemic progression after stereotactic radiotherapy with a median PFS of 9.6 months; subsequent SBRT after disease progression was applied in 8 patients (Fig. 1b).

Out of 48 lesions treated with stereotactic radiotherapy, 5 patients experienced local failure. The local control rates at 6, 12, and 36 months were 91.3, 87.7, and 83%, respectively (Fig. 1c).

For patients who had not received any systemic treatment after being diagnosed with the metastatic disease and who had received stereotactic radiotherapy first (de novo OMD), the median STFS was 23.9 months (Fig. 1d).

A univariate analysis was conducted using the log-rank test to investigate the prognostic factors influencing the survival outcomes for OMD/R-HNC patients. Synchronous and metachronous OMD showed no significant difference regarding OS or PFS (p-value: 0.67 and 0.73, respectively; Fig. 2a). In addition, squamous cell cancer and nonsquamous cell cancer histologies did not influence OS and PFS (Fig. 2b). However, we observed a trend for better OS in patients with p16-positive tumors compared to p16-negative tumors, specifically among those with a squamous cell cancer histology (p-value 0.074; Fig. 2c). We also examined the impact of the time to progression (TTP) of OMD from the date of diagnosis using the time-dependent ROC curve analysis between the TTP and OS. We found that a cut-point of TTP < 150 days showed the most robust significant survival difference. Patients who progressed within 150 days after diagnosis (TTP < 150 days) had, besides a lower PFS, also a significantly lower OS (p-value < 0.001; Fig. 2d). In addition, a de novo OMD showed a significantly better PFS than the induced OMD (p-value = 0.025). However, neither had a meaningful statistical difference in OS (p-value = 0.63; Fig. 2e). Understandably, patients who received chemotherapeutics (including cetuximab) throughout their metastatic disease had a significantly lower PFS (p-value = 0.025). However, the OS difference with those not receiving the chemotherapeutics was statistically nonsignificant (p-value = 0.34; Fig. 2f). Furthermore, patients who received immunotherapy (anti-PD-1) throughout the treatment of their metastatic disease did not show a significant improvement for either PFS or OS (p-value = 0.25 and 0.31, respectively; Fig. 2g).

There was no statistical difference in local control between treated pulmonary and non-pulmonary lesions (p-value = 0.93), nor was there a statistically significant difference based on the histology of the lesion (squamous vs. nonsquamous; p-value = 0.48). ROC analysis showed the greatest impact on local control was at a dose cut-point of EQD2 at 87.25 Gy. However, there was no statistically significant difference in local control between doses lower or higher than 87.25 Gy (p-value = 0.17).

Using Cox regression for the multivariate analysis, p16-positive squamous cell cancer, metachronous OMD, and TTP are positive predictors for OS (Fig. 3a). Furthermore, de novo OMD was the only positive predictor in the Cox regression model for PFS (Fig. 3b). Lastly, radiation dose (EQD2), histology, and p16-status were not predictors for local control (Fig. 3c).

Two patients (6.5%) suffered from grade 3 dysphagia: 1 patient after SBRT of regional recurrence and 1 patient after repeated SBRT of mediastinal metastatic lymph nodes. No grade 5 toxicity was documented from the treatment.