Prognostic Value of Radiological Extranodal Extension Detected by Computed Tomography for Predicting Outcomes in Patients With Locally Advanced Head and Neck Squamous Cell Cancer Treated With Radical Concurrent Chemoradiotherapy

Objective Extra Nodal Extension (ENE) assessment in locally advanced head and neck cancers (LAHNCC) treated with concurrent chemo radiotherapy (CCRT) is challenging and hence the American Joint Committee on Cancer (AJCC) N staging. We hypothesized that radiology-based ENE (rENE) may directly impact outcomes in LAHNSCC treated with radical CCRT. Materials and Methods Open-label, investigator-initiated, randomized controlled trial (RCT) (2012–2018), which included LAHNSCC planned for CCRT. Patients were randomized 1:1 to radical radiotherapy (66–70 grays) with concurrent weekly cisplatin (30 mg/m2) [cisplatin radiation arm (CRT)] or same schedule of CRT with weekly nimotuzumab (200 mg) [nimotuzumab plus CRT (NCRT)]. A total of 536 patients were accrued and 182 were excluded due to the non-availability of Digital Imaging and Communications in Medicine (DICOM) computed tomography (CT) data. A total of 354 patients were analyzed for rENE. Metastatic nodes were evaluated based on five criteria and further classified as rENE as positive/negative based on three-criteria capsule irregularity with fat stranding, fat invasion, and muscle/vessel invasion. We evaluated the association of rENE and disease-free survival (DFS), loco-regional recurrence-free survival (LRRFS), and overall survival (OS). Results A total of 244 (68.9%) patients had radiologically metastatic nodes (rN), out of which 140 (57.3%) had rENE. Distribution of rENE was balanced in the two study groups CRT or NCRT (p-value 0.412). The median follow-up period was 39 months (ranging from 35.5 to 42.8 months). Complete response (CR) was seen in 204 (57.6%); incomplete response (IR), i.e., partial response plus stable disease (PR + SD), in 126 (35.6%); and progressive disease (PD) in 24 (6.8%). rENE-positive group had poor survival compared to rENE-negative group 3-year OS (46.7% vs. 63.6%), poor DFS (48.8% vs. 87%), and LRRFS (39.9% vs. 60.4%). rENE positive had 1.71 times increased risk of IR than rENE negative. Overall stage, site, clinical metastatic node (cN), response, and rENE were the significant factors for predicting OS, DFS, and LRRFS on univariate analysis. After making adjustment on multivariate analysis, rENE was an independent prognostic factor for DFS and trending to be significant for OS. Conclusion Pre-treatment rENE is an independent prognostic marker for survival in patients with LAHNSCC treated radically with CCRT that can be used as a potential predictive marker for response to treatment and hence stratify patients into responders vs. non-responders. We propose the mahajan rENE grading system applicable on CT, magnetic resonance imaging, positron emission tomography–contrast-enhanced CT, and ultrasound.


INTRODUCTION
The commonest cancer histology in the head and neck is squamous cell carcinoma (SCC) (1). CCRT is a standard treatment option in LAHNSCC. It serves either as definitive treatment in cancers of oropharynx, larynx, and hypopharynx or as adjuvant treatment for oral cavity cancers post-surgery in presence of pathological ENE (pENE) or positive tumor margins on histopathology (2,3).
ENE is an important prognostic and predictive factor that has been accorded a place in American Joint Committee on Cancer (AJCC) eight edition N category for head and neck cancers and has become the criteria for N3b disease (4). pENE has been extensively studied, and it can be microscopic (ENEmi) or macroscopic (ENEma). A study by Tirelli et al. found that the 3-year OS was 46% in the ENEmi group and 38.9% in the ENEma group (5). Another study by Thomas et al. found that pENE and clinical ENE (cENE) were associated with 60% decrease in 5-year OS compared to negative ENE (6). However, as ENE is pathologically assessed, this important information is not available for patients who are radically treated with definitive chemoradiotherapy and histopathological ENE is not available. Hence, imaging-based assessment of ENE [radiology-based ENE (rENE)] using various imaging criteria can help in better prognostication and planning with the intensification of adjuvant treatment in LAHNSCC. Few recent studies have shown the clinical implications of rENE in head and neck cancers (7). A study by Benjamin et al. found that rENE-positive patients had significantly worse 3-year OS (95% vs. 77%), progression-free survival (91% vs. 71%), and distant control (98% vs. 81%) than rENE-negative patients in locally advanced oropharyngeal carcinoma (8).
Multiple imaging-based studies have been performed in the past to know the diagnostic accuracy of rENE using CT, magnetic resonance imaging (MRI), and ultrasonography (US) and correlating it with pENE. Few studies have compared the diagnostic accuracy of different imaging modalities; for example, the meta-analysis by Su et al. (9) reported comparable results between CT and MRI in predicting ENE with sensitivity and specificity of 77% and 85% for CT and 85% and 84% for MRI, respectively. They also found sensitivity of 86% and specificity of 86% for positron emission tomography-computed tomography (PET-CT) and of 87% and 75%, respectively, for ultrasound (US). Almulla et al. showed that compared to MRI, CT showed improved sensitivity, negative predictive value, and diagnostic accuracy but similar specificity and positive predictive value (10). CT has been found reliable for evaluating rENE, with sensitivity ranging from 64% to 73% and specificity ranging from 82% to 87%, respectively. A study by Url et al. found sensitivity of CT in rENE detection of 73% and specificity of 91% (11). Given the high specificity and negative predictive value, it is worthwhile to explore implications of rENE in patients treated radically with CCRT, where pENE is not available. CT-based rENE has the potential to be used to accurately stage by predicting ENE and to stratify patients into high and low risk before planning the radical treatment in LAHNSCC.
With this background, we hypothesize that rENE and hence the N category based on it may have a potential to be prognostic and predictive marker in determining outcomes, we performed this analysis on the dataset available from already published randomized controlled trial (CTRI/2014/09/004980) (2). The present study is a post hoc analysis that evaluates the impact of rENE on survival outcome on a cohort of patients who were prospectively followed up.

Materials
The study was approved by institutional ethics committee and was registered with clinical trial registry of India (CTRI/2014/09/ 004980). All patients underwent standard study protocol for evaluation and treatment. This was an investigator-initiated, randomized controlled trial, which allocated LAHNSCC (AJCC seventh edition stage III and IV oropharyngeal, laryngeal, and hypopharyngeal cancer) patients in ratio of 1:1 to receive either radical radiotherapy (66-70 Gy) with concurrent weekly cisplatin (30 mg/m 2 ) (CRT) or the same schedule of CRT with weekly nimotuzumab (200 mg) (NCRT). The stratified block randomization was performed and trial population was stratified by five factors, namely, site of malignancy (oropharynx versus others), stage (III versus IV), age (≤60 versus >60 years), radiotherapy technique, and treatment center. The key inclusion criteria for the current analysis were treatment-naive oropharyngeal, laryngeal, and hypopharyngeal patients with pretreatment contrast-enhanced CT (CECT)/PET-CT available in our institute. Patients who had received any treatment prior, unavailable DICOM data on imaging or who presented with distant metastasis, inadequate treatment details, and follow-up data were excluded. Detailed inclusion and exclusion criteria of the trial are mentioned in the supplement of the main published article of the trial (2). A total 536 patients were enrolled in this trial between 2012 and 2018. Out of these, 182 patients were excluded from this analysis due to the exclusion criteria mentioned above. The final analysis was performed on a cohort of 354 patients: 181 oro-pharyngeal and 173 other sites (larynx and hypopharynx) cancers. The consort diagram is shown in Figure 1.

Treatment Protocol and Follow-Up
Both the arms were administered high-dose, curative radiotherapy for 6.5 to 7 weeks using a standard twodimensional (2D) technique, a three-dimensional (3D) conformal technique, or intensity-modulated radiotherapy with megavoltage radiation. Local tumor and lymph node disease were treated with 70 grays (Gy), in 2 Gy per fraction, 5 days per week, whereas upto dose of 46 to 50 Gy was planned for uninvolved nodal regions of the neck. Other altered fractionation schedules were permitted if the biologic equivalent dose for tumor control was similar to 70 Gy at 2 Gy per fraction. Cisplatin was given at a dose of 30 mg/m 2 weekly during radiation along with supportive medication in both arms. Weekly Nimotuzumab in addition was given in the NCRT arm intravenously as a 200-mg dose in 250 ml of normal saline over 60 minutes without any premedication. The response assessment of PET-CECT was performed 8 weeks after the completion of active treatment. The baseline imaging was performed prior to the initiation of treatment and subsequent imaging performed as per the schedule, which was compared for Response Evaluation Criteria in Solid Tumours (RECIST). The patients were followed up clinically and with CECT at 12 and then 24 weeks from randomization for initial 6 months. Subsequently, patients underwent clinicoradiological evaluation every 6 months for the next 2 years. After the completion of 2 years, patients were followed up only clinically until the completion of 5 years. Regional recurrences that were detected on clinical examination or imaging were confirmed with fine needle aspiration cytology.

Image Analysis
The baseline imaging data (CECT images) of the eligible patients were reviewed by two dedicated head and neck radiologists (AM and NS with 10 and 7 years of experience, respectively) independently. Nodes were categorized into metastatic and non-metastatic and further into rENE positive and rENE negative. The criteria used for identifying metastatic node included round shape, loss of fatty hilum, necrosis, heterogeneous enhancement, and irregular capsule ( Figure 2). Nodes with the presence of any two or more criteria out of these five criteria were considered positive for nodal metastasis to improve the specificity. The criteria for rENE included capsular irregularity with fat stranding, capsule irregularity with fat invasion and gross muscle/vessel invasion ( Figure 3). rENE was considered positive if any one criterion was present. The final metastatic and rENE was reported patientwise, i.e., patients were defined as node and rENE positive and negative. The individual features were reported independently by both the radiologist, and if there was discordance between the final status of nodal metastasis or rENE, then both the radiologists reviewed the scans together and reached a consensus. We evaluated the association of rENE and clinical outcomes.

Statistical Analysis
The patients' demographics, treatment, and outcome data were entered in IBM SPSS version 21.0. The Chi-square test and Fisher's exact test were used to analyze descriptive data. The value of p < 0.05 was considered statistically significant. Kaplan-Meier estimates were performed for survival analysis, and comparison was done using log-rank test. The OS was calculated from the date of randomization to the date of death. DFS was calculated from the time when the patient was diseasefree, i.e., if the patient was disease-free (achieved complete remission) after chemoradiotherapy, then it was calculated from the date of response assessment and in patients of residual disease that underwent salvage surgery; it was calculated from the date of salvage surgery to the date of any locoregional or distant disease recurrence. In all other patients who were never disease-free, it was considered 0. LRRFS was calculated from the date of randomization to the date of local and/or regional recurrence. Univariate cox analysis was performed to calculate hazard ratio for potential prognostic factors, including age at diagnosis, gender, site of tumor, stage of tumor (cT category, cN category, and overall stage), response, rENE, cumulative cisplatin dose >200 mg/m 2 , and treatment regime. Statistically significant factors on univariate analysis and clinically important covariates were used to create a multivariate Cox proportional hazards regression model to determine whether rENE is an independent prognostic factor.

OS, DFS, and LRRFS Univariate Analysis
The Kaplan-Meier curve ( Figures 5-7) showed 3-year OS, DFS, and LRRFS for the hypopharynx-larynx were significantly higher as compared to oropharynx (p < 0.001). Patients with the clinically node positive disease had poor survival than the clinically node negative disease (p = 0.001). Stage IV had poor survival than stage III patients (p < 0.001). When the response was assessed according to RECIST criteria, CR had better survival than PR, and PD showed worst survival (p < 0.001). When compared to rENE status, patients with rENE-positive status had poor survival [OS (p < 0.001), DFS (p < 0.001), and LRRFS (p = 0.003)] compared to rENE-negative patients ( Figures 5-7). On performing the cox univariate analysis gender, site, cN, stage, rENE, and response had a significant impact on OS (Table 3). Site, cN, stage, rENE, response, and treatment arm were significant prognostic factors for DFS and LRRFS (Table 3).

OS, DFS, and LRRFS Multivariate Analysis
On multivariate analysis, the independent prognostic factors for OS were site (oropharynx vs. laryngo-hypopharynx, p = 0.004) and stage (stage III vs. stage IV, p = 0.015). The presence of rENE had a trend toward inferior OS (p = 0.069) ( Table 4). rENE was an independent prognostic factor of DFS (p = 0.021) along with the site of disease (p = 0.017) ( Table 4). Only site and the treatment arm (CRT vs. NCRT) were independent prognostic factors for LRRFS ( Table 4).

Factors Predicting Clinical Response
Site, cN, and rENE were significant factors for predicting the response to treatment both on univariate and multivariate analysis (Tables 5, 6). rENE-positive patients had 1.71 times increased risk of IR than negative patients.

DISCUSSION
Extranodal extension/extracapsular spread (ENE/ECS) is one of the most important independent prognostic factors in head and neck SCC (HNSCC) and has been found to show a direct impact on poor survival outcomes (12,13). pENE has been extensively studied in the past and is considered to be the gold standard for deciding adjuvant treatment in HNSCC (12). Addition of chemotherapy to radiotherapy as adjuvant (CCRT) improves the 5-year survival rate compared to RT alone (14). In patients       with LAHNSCC who do not undergo surgery, precise assessment of rENE is of critical importance for planning the appropriate treatment, for prognosticating the disease, and for predicting treatment response. This study aimed to assess the prognostic importance of rENE detected by CT in predicting clinical response after CCRT and its implication on survival outcomes. According to the results of our study, rENE was an independent prognostic factor for DFS and showed trend toward significance for OS. A recent study by Moon et al. (7) showed similar results where rENE (+) had worse 5-year OS (74% vs. 94%, p < 0.01) and DSS (42% vs. 84%, p < 0.01) compared with the rENE (−) cohort. The other important factors affecting survival were the site of the disease, stage, and treatment arm, which are similar to the previously reported articles (2).
It is of paramount importance to accurately stage the nodal disease in patients radically treated with CTRT, which is based on clinical and radiological findings, due to its recent incorporation in the AJCC staging eight edition. It is necessary for appropriate radiation planning and intensification of treatment. Failure to do so will result in residual/recurrent disease with poorer outcomes (15,16). We found the presence of rENE, cN, and site of primary to be an independent prognostic factor in deciding clinical response. rENE-positive patients had 1.71 times increased risk of IR than negative patients. Patients with cN had two times more chances of PR compared to node-negative patients.
In the eighth edition of the AJCC TNM classification, cENE was included for N3b stage classification and risk stratification of HNSCC because of uncertainty about the reliability of rENE (4). However, multiple studies now suggest that rENE can be reliably ascertained with high specificity if strict imaging criteria are used (17)(18)(19)(20)(21)(22)(23). Many previous reports have compared concordance between rENE and pENE (17)(18)(19)(20)(21)(22). It has the potential to refine the cN classification and facilitate treatment selection in both viral-related and viral-unrelated HNSCC (24)(25)(26)(27). Diagnosis of rENE based on our criteria has shown it to be a good prognostic factor for survival and predictor of response to treatment. The assessment of rENE with improved diagnostic accuracy is most important to reduce false-positive and false-negative results. There is a lack of consensus regarding the imaging criteria for  assessing ENE from the past studies (28)(29)(30)(31). Multiple imaging criteria were used by different studies ( Table 7); for example, example Url et al. proposed apparent fat and soft tissue infiltration and infiltration of muscle and carotid sheath (11). A study by Aiken et al. (20) found a strong association between radiologically determined lymph node necrosis and pathological extracapsular spread (pENE) (p<0.01), a similar result showed by Randall et al. (16). A study by Carlton et al. found 48% sensitivity and 86% specificity for capsular irregularity imaging criteria (17). A recent study by Faraji et al. concluded that of the seven imaging features hypothesized to be associated with ENEstatus, the presence of irregular nodal margins and absence of perinodal fat plane were the most specific and sensitive features, respectively (22). The study by both Randall et al. and Aiken et al. found that central necrosis was the best predictor of ENE among multiple other imaging criteria. We found necrosis in 74% of rENE positive cases. Prabhu et al. found that sensitivity ofCT detection of ECS was 18% when the extent of histopathological ENE from the capsule was less than or equal to 1 mm and that sensitivity and accuracy improved as the extent increased more than 1 mm. With this in mind, we propose the imaging-based mahajan ENE grading system, which will be applicable on CT, MRI, PET-CECT, and US. The rENE grading system with its cENE and pENE correlates is shown in Figure 8.
There were few limitations of our study; as patients were radically treated with CCRT and only a few underwent salvage neck dissection, rENE status confirmation was not available. We also could not study the diagnosis and prognostic implication of microscopic versus macroscopic rENE. Indian population has predominance of human papilloma virus (HPV)-negative oropharyngeal cancers. Same was evident from our trial as well; of the 269 patients with oropharyngeal cancer, p16 was feasible only in 212 patients. In addition, 187 of 212 patients were p16 negative, and the analysis of the same has already been published (32). Therefore, it was not possible to study the impact of rENE in HPV-positive oropharyngeal cancers, which is another research avenue.
To conclude, pre-treatment rENE is an independent prognostic marker for survival in patients with LAHNSCC treated radically with CCRT that can be used as a potential predictive marker for response to treatment and hence stratify patients into responders vs. non-responders. We propose the mahajan rENE grading system applicable on CT, MRI, PET-CECT, and US.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by IEC TMC. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.  the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

AUTHOR CONTRIBUTIONS
Copyright © 2022 Mahajan, Chand, Agarwal, Patil, Vaish, Noronha, Joshi, Kapoor, Sable, Ahuja, Shukla, Menon, Agarwal, Laskar, D' Cruz, Chaturvedi, Chaukar, Pai, Pantvaidya, Thiagarajan, Rane and Prabhash. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.