Recently, ICG-guided surgery has gained in popularity, exemplified by Xia et al. [7] who reported a sensitivity of 98.1% and specificity of 89.1% in the detection of metastatic nodes. In addition, a recent study [8] demonstrated the ability of ICG to identified neck metastases in a population of young adults and children with a sensitivity of 83%, a specificity of 88%, a positive predictive value (PPV) of 91%, and a negative predictive value (NPV) of 86%. Another recent study reported a sensitivity of 100% for pathologic nodes during neck dissection [9]. However, all of these studies were performed in malignancies that were not treated previously with radiation therapy.

Locally advanced HNSCC mainly requires multimodal treatment and RT, with or without chemotherapy, is often the first therapeutic option proposed to these patients [10, 11]. However, the aggressive nature of these cancers often leads to recurrence and, in this setting, surgery remains the only curative option [12]. As compared with first-line surgery, salvage interventions after RT are associated with higher rates of morbidity [13]. For neck dissection, in particular, RIF can impair visual and palpatory assessment, making it more difficult to distinguish between healthy, fibrotic, and neoplastic tissue.

In a previous study [14] in a heterogeneous population in terms of primary head and neck tumours and previous RT, we evaluated the feasibility of near infrared (NIR) fluorescent imaging after intravenous injection of ICG during neck dissection. We found that the presence of a fluorescent node was associated with a 14.1-fold risk of invasion regardless of the size of the node. In this feasibility study, fluorescence magnitude was correlated to node invasion and was able to distinguish between healthy and invading nodes after IV injection. However, we acknowledged that larger and more homogenous series were required to define the optimal role of NIR fluorescence in head and neck cancer and its potential routine utilization. Indeed, in this study, samples from one patient did not exhibit node fluorescence after IV injection of ICG. This patient underwent two sessions of radiotherapy and one surgery with node dissection before the last procedure with fluorescence imaging. For this reason, we decided to perform an objective measurement of a homogeneous population of patients with HNSCC with two comparable subgroups (with or without previous RT). However, more irradiated patients were included in the study since it has been established that the yield of lymph nodes is decreased in irradiated patients compared to non-irradiated patients [15].

To the best of our knowledge, the question of the influence of radiation therapy on node fluorescence in a population of patients with HNSCC has not been addressed yet. However, a recent study [16] assessed, for the first time, the influence of radiotherapy on the fluorescence of primary HNSCC after IV injection of ICG. Four patients with primary tumour recurrence after radiotherapy were included. Two patients had subjectively increased fluorescence compared to surrounding tissue. The last two patients had a tumour developed on a bed of lichen, one had moderately increased fluorescence compared to adjacent tissue and the last patient had no difference in fluorescence between the tumour and healthy tissue. The authors concluded that NIR fluorescence mapping in HNSCC patients previously subjected to radical radiotherapy clearly established the feasibility of using this technique to delineate tumours from lichen.

Based on these observations, it could be suspected that specific limitations of this fluorescence imaging technique could be encountered in patients who have been previously irradiated in the same region. Accordingly, the objective of the present work was to prospectively verify the feasibility of fluorescence imaging to detect LNs associated with HNSCC in patients who had undergone previous RT, as compared with those who did not receive RT before surgery.

Our main observation was that fluorescent LNs could be identified in patients who were previously irradiated, and regarding the total number of harvested nodes in the irradiated and non-irradiated populations, we did not observe a difference in fluorescence values (p = 0.63). Furthermore, we did not observe a significant difference between irradiated and non-irradiated metastatic nodes (p = 0.23). However, the average fluorescence of non-irradiated metastatic lymph nodes was 28.1 AU vs. 19AU in irradiated patients, suggesting a trend towards a decreased fluorescence in metastatic nodes in irradiated patients.

To better understand the factors influencing the value of fluorescence, we correlated the size of the metastatic node and the size of the metastatic component to the value of fluorescence. We found that the surface of the metastatic node and the metastatic component were correlated to the fluorescence value (p = 0.006 and p = 0.007, respectively), confirming our previous observations [14].

However, regarding the subgroup of irradiated patients, we did not find a correlation between the surface of the metastatic node and the fluorescence. This observation could be explained by necrotic component in irradiated nodes. In our serie, histopathological reports described necrotic component among nodes in 3 irradiated patients.

After adjustment for a previous RT, the correlation between the surface of a metastatic node and the size of the metastatic component remained significant (p = 0.02). This result suggests that the value of fluorescence is correlated to the amount of neoplastic tissue inside the invaded node. Therefore, the technique will probably be ineffective for the identification of micrometastases. However recent studies have shown in non-irradiated patients an in vivo sensibility of 100% in the identification of nodes metastasis and the identification of nodes metastasis outside the planned resection area [17, 18].

Thus, the main clinical application of the technic would probably be the identification of nodes metastasis outside the planned neck dissection area.