Abstract
Background and purpose
In patients with squamous cell carcinoma of the head and neck (HNSCC), extracapsular spread (ECS) of metastases in cervical lymph nodes affects prognosis and therapy. We assessed the accuracy of intravenous contrast-enhanced computed tomography (CT) and the utility of imaging criteria for preoperative detection of ECS in metastatic cervical lymph nodes in patients with HNSCC.
Materials and methods
Preoperative intravenous contrast-enhanced neck CT images of 93 patients with histopathological HNSCC metastatic nodes were retrospectively assessed by two neuroradiologists for ECS status and ECS imaging criteria. Radiological assessments were compared with histopathological assessments of neck dissection specimens, and interobserver agreement of ECS status and ECS imaging criteria were measured.
Results
Sensitivity, specificity, positive predictive value, and accuracy for overall ECS assessment were 57%, 81%, 82% and 67% for observer 1, and 66%, 76%, 80% and 70% for observer 2, respectively. Correlating three or more ECS imaging criteria with histopathological ECS increased specificity and positive predictive value, but decreased sensitivity and accuracy. Interobserver agreement for overall ECS assessment demonstrated a kappa of 0.59. Central necrosis had the highest kappa of 0.74.
Conclusion
CT has moderate specificity for ECS assessment in HNSCC metastatic cervical nodes. Identifying three or more ECS imaging criteria raises specificity and positive predictive value, therefore preoperative identification of multiple criteria may be clinically useful. Interobserver agreement is moderate for overall ECS assessment, substantial for central necrosis. Other ECS CT criteria had moderate agreement at best and therefore should not be used individually as criteria for detecting ECS by CT.
Keywords: Extracapsular extension, extracapsular spread, head and neck, squamous
Introduction
Extracapsular spread (ECS) occurs when metastatic tumor cells within a lymph node breech the lymph node capsule. In patients with squamous cell carcinoma of the head and neck (HNSCC), the presence of ECS in metastatic cervical lymph nodes portends a poor prognosis. When histopathological ECS is identified, the 5-year mortality rate is reduced to 27%, compared to 70% when ECS is absent.1 ECS is correlated with increased rates of recurrence, contralateral recurrence, occurrence in unusual sites, and distant metastasis.2–6 ECS is identified in up to 50% of neck dissection specimens and in approximately 20% of patients with HNSCC who have clinically negative neck examinations.7 The current standard for the identification of ECS in patients with HNSCC is histopathological examination of neck dissection specimens. The extent of ECS by histopathological assessment, whether greater or less than 2 mm, does not have a significant impact on patient outcomes.8,9
The 2016 National Comprehensive Cancer Network head and neck cancer guidelines include computed tomography (CT) and/or magnetic resonance imaging (MRI) for initial diagnostic imaging in head and neck cancer.10 CT has been found to be slightly better than MRI for neck staging purposes in head and neck cancers.11 When cervical metastasis is suspected due to physical examination or imaging findings, patients often undergo primary tumor excision as well as neck dissection. The specific type of neck dissection, whether comprehensive, selective, or elective, and the laterality of dissection, is largely dependent on preoperative imaging, as well as the primary echelon drainage pathway of the tumor. Decisions regarding dissection approach are multidisciplinary involving surgery, radiation oncology, medical oncology, radiology, and the patient or their representative. Following excision, adjuvant chemoradiation therapy has been shown to improve overall survival compared to adjuvant radiation alone for patients with HNSCC and ECS identified in neck dissection specimens.12 Preoperative awareness of disease extent, including ECS status, could allow us to determine candidacy for definitive chemoradiotherapy or surgery.
Evidence for the diagnostic accuracy of imaging for ECS continues to develop. Recent studies have demonstrated a range of sensitivities between 49% and 91% and specificities from 50% to greater than 90%.13–21 These wide-ranging results represent the variable criteria and approaches utilized for radiological assessment for ECS in the HNSCC metastatic cervical nodes. Our objectives were to perform a retrospective observational study to determine the sensitivity and specificity of CT, and the utility of employing imaging criteria on CT to identify ECS of cervical lymph node metastases from HNSCC.
Materials and methods
Data collection
The internal review board approved this study and granted exemption from informed consent. Records of every neck dissection performed at our tertiary center were obtained between 2007 and 2015 by reviewing billing records for neck dissections. The patient charts for each neck dissection were reviewed in reverse chronology and a list of subjects was generated that met our inclusion criteria and exclusion criteria. Inclusion criteria were that patients had primary non-recurrent HNSCC with cervical node metastases identified by histopathological examination of neck dissection specimens. Also, patients must have had an intravenous contrast CT of the neck performed within 8 weeks of neck dissection. Exclusion criteria were that subjects could have no prior history of neck surgery at the site of dissection, no prior history of head or neck radiation, and no prior history of chemotherapy. Contrast-enhanced CT images were archived in the University of Missouri picture archiving and communication system (PACS). Histopathological examinations of neck dissection specimens were performed by board-certified University of Missouri pathologists. The clinical data were obtained from patients’ electronic records. Data reviewed and collected included patient demographics, medical history, imaging reports, intraoperative findings, tumor board discussions, and pathology reports.
Imaging methods
Patients underwent imaging on one of several commercially available CT systems with multidetector capability. Onsite imaging CT studies were performed on a Somatom Definition 64-slice dual source (Siemens, Erlangen, Germany). Our split-bolus technique used a total of 100 mL of intravenous Visipaque (iodixanol injection 320 mg/ml; GE Healthcare Inc., Princeton, NJ, USA) injected at a rate of 2.5–3 cc/second, followed by a 30-second delay. Injection protocol included a precontrast and postcontrast saline bolus of 20 cc and 30 cc, respectively, at a rate of 2.5 cc/second. We acquired contiguous axial images from the external auditory meatus to the top of the aortic arch with the following settings: section thickness, 3 mm; collimation, 0.6 mm; pitch, 0.8:1; gantry rotation time, 1 second; field of vision, 25 cm; 120 kVP. Reformatted images with 3-mm section thicknesses in the axial planes and 3-mm sagittal and coronal reformations were sent to a PACS.
Imaging review
Two neuroradiologists specialized in head and neck imaging, both with more than 5 years of post-fellowship experience, assessed for overall ECS assessment and the presence or absence of five ECS imaging criteria. ECS imaging criteria included irregular nodal enhancement, infiltration into adjacent tissues, indistinct nodal margins, matted nodes, and central necrosis. Matted nodes was assessed when three abutting nodes had lost the intervening fat plane, presumably to have been replaced by ECS. Our observers assessed every conspicuous node using the above criteria, irrespective of node size. The list of CT studies for review were randomly assigned by date and provided to the observers. Observers separately reviewed the intravenous contrast-enhanced CT study most recent to the subject’s neck dissection date and provided individual impressions. Observers were aware that all subjects had at least one node positive for metastatic disease, but the observers were blinded to ECS status.
Statistical analysis
The statistical methods used include estimates of proportions for individual and consensus sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy with exact confidence intervals used due to small sample sizes. Cohen weighted kappa statistics were calculated for evaluating interobserver agreement. Chi-square tests were used to examine the relationship between ECS criteria and histopathological ECS. The statistical software SAS version 9 (SAS Institute Inc., Cary, NC, USA) was used for analysis.
Results
Intravenous contrast-enhanced neck CT images of 93 patients with metastatic HNSCC were reviewed separately by the two observers. Table 1 represents patient demographics, primary tumor sites, dissection laterality, and histopathology stage. Fifty-six cases had ECS identified during histopathological examination of neck dissection specimens, equating to 60% ECS for our sample of HNSCC patients. The average age of the patient population was 61 years. Most patients were white men and most patients had a history of alcohol and tobacco use. The majority of specimens were of unilateral neck dissections, while 26 were bilateral. Less than five of the dissections were comprehensive. There was a variety of primary tumor sites for the squamous cell carcinoma, with primary tumor sites at the oral cavity, oropharyngeal, and laryngeal representing the majority of cases. Over half of the oropharyngeal carcinomas were human papilloma virus positive. Tumor stages were widely variable as well.
Table 1.
Demographics and disease characteristics.
| Parameter | Value |
|---|---|
| Mean age (years) | 61 ± 11 SD |
| Gender (M/F) | 58/25 |
| Race (white/non-white) | 88/5 |
| Alcohol use | 56 |
| Tobacco use | 78 |
| Primary tumor site | |
| Oral cavity | 37 (3 HPV+) |
| Oropharynx | 17 (9 HPV+) |
| Larynx | 14 (2 HPV+) |
| Mandible | 8 |
| Lip | 3 |
| Nasopharynx | 2 |
| Submandibular gland | 1 (1 HPV+) |
| External carotid, auricle, parotid, hypopharynx | 1 of each |
| Unspecified location | 7 |
| HP stage | |
| N1 | 31 |
| N2a | 5 |
| N2b | 32 |
| N2c | 20 |
| N3 | 5 |
| HP ECS assessment | |
| ECS present/absent | 56/37 |
ECS: extracapsular spread; HP: histopathological; HPV: human papilloma virus.
Table 2 shows the individual and consensus sensitivities, specificities, PPV, NPV, and accuracy of overall ECS assessment and for when three or more imaging criteria were identified. Observer 1 assessed 30 of 37 histopathological negative cases correctly, and 32 of 56 histopathological positive cases correctly. Observer 2 assessed correctly 28 of 37 histopathological negative cases, and assessed 37 of 56 positive cases correctly. Observer 1 identified three or more imaging criteria in 27 cases, and 23 of those 27 cases were ECS positive per histopathology report. Observer 2 identified the presence of three or more imaging criteria in 35 cases, 30 of which were ECS positive per histopathology report. The consensus assessment represents only the cases in which both radiologists agreed. Consensus agreement occurred in 80% of the cases, 74 out of 93. Consensus slightly improved accuracy. Figure 1 represents CT images reviewed by observers and addresses their assessments in correlation with ECS status according to histopathological evaluation.
Table 2.
Test.
| Sensitivity in % (95% CI) | Specificity in % (95% CI) | PPV in % (95% CI) | NPV in % (95% CI) | Accuracy | |
|---|---|---|---|---|---|
| Observer 1 ECS assessment | 57 (43–70) | 81 (65–92) | 82 (67–91) | 56 (42–68) | 67 |
| Observer 1 3+ criteria identified | 41 (28–55) | 89 (73–96) | 85 (68–94) | 50(38–61) | 60 |
| Observer 2 ECS assessment | 66 (52–78) | 76 (59–88) | 80 (67–89) | 60 (45–72) | 70 |
| Observer 2 3+ criteria identified | 54 (40–67) | 86(70–95) | 86 (69–95) | 55 (42–68) | 67 |
| Consensus ECS assessment | 65 (49–79) | 84 (66–95) | 85 (69–93) | 63 (48–76) | 72 |
ECS: extracapsular spread; CI: confidence interval; PPV: positive predictive value; NPV: negative predictive value.
Figure 1.
Arrows point to nodes of interest. (a) True +: Level 1 b node positive for HNSCC and ECS; observers identify infiltration into adjacent tissues, indistinct nodal margins, and central node necrosis. (b) True −: Level 1 b node positive for HNSCC but no ECS; observers do not identify any ECS criteria. (c) False −: Level 2 positive for 1 cm HNSCC focus and ECS; observers do not identify any ECS criteria. (d) False +: Level 2 node positive for HNSCC but no ECS; observers identify irregular nodal enhancement, infiltration into adjacent tissues, indistinct nodal margins, matted nodes and central necrosis. HNSCC: squamous cell carcinoma of the head and neck; ECS: extracapsular spread.
Interobserver agreement regarding overall assessment and individual imaging criteria is represented in Table 3, in which Cohen’s kappa was used to measure agreement between the two observers. Interobserver agreement for overall ECS assessment was moderate with a kappa of 0.59. Central necrosis had the highest agreement of the imaging criteria, with a kappa of 0.74, while the other imaging criteria had moderate agreement at best. Tables 4 and 5 show the sensitivities and specificities of each individual criterion for identifying ECS alone, and none of the individual criteria had substantial accuracy independently.
Table 3.
Interrater agreement criteria.
| Kappa | |
|---|---|
| Radiological ECS assessment | 0.59 (0.43–0.75) |
| Indistinct nodal margins | 0.10 (0.3–0.23) |
| Infiltration into adjacent tissues | 0.51 (0.33–0.69) |
| Irregular nodal enhancement | 0.37 (0.19–0.55) |
| Matted nodes | 0.50 (0.28–0.72) |
| Central necrosis | 0.74 (0.61–0.88) |
ECS: extracapsular spread.
Table 4.
Observer 1 criteria identified.
| Sensitivity in % (95% CI) | Specificity in % (95% CI) | PPV in % (95%CI) | NPV in % (95%CI) | Accuracy | |
|---|---|---|---|---|---|
| Indistinct nodal margins | 7 (2–17) | 100 (91–100) | 100 (51–100) | 42 (32–52) | 44 |
| Infiltration into adjacent tissues | 50 (36–64) | 84 (68–94) | 82 (67–92) | 53 (40–65) | 63 |
| Irregular nodal enhancement | 34 (22–48) | 95 (82–99) | 90 (71–97) | 47 (37–60) | 58 |
| Matted nodes | 27 (16–40) | 86 (71–95) | 75 (53–89) | 44 (33–55) | 51 |
| Central necrosis | 63 (49–75) | 78 (62–90) | 81 (67–90) | 58 (44–71) | 69 |
CI: confidence interval; PPV: positive predictive value; NPV: negative predictive value.
Table 5.
Observer 2 criteria identified.
| Sensitivity in % (95 % CI) | Specificity in % (95 % CI) | PPV in % (95%CI) | NPV in % (95%CI) | Accuracy | |
|---|---|---|---|---|---|
| Indistinct nodal margins | 48 (35–62) | 86 (71–95) | 84 (68–93) | 52 (40–64) | 63 |
| Infiltration into adjacent tissues | 50 (36–64) | 86 (71–95) | 85 (69–93) | 53 (41–65) | 65 |
| Irregular nodal enhancement | 61 (47–74) | 84 (68–94) | 85 (71–93) | 58 (45–71) | 70 |
| Matted nodes | 20 (10–32) | 92 (78–98) | 79 (52–92) | 43 (33–54) | 48 |
| Central necrosis | 64 (50–77) | 70 (53–84) | 77 (63–86) | 57 (42–68) | 67 |
CI: confidence interval; PPV: positive predictive value; NPV: negative predictive value.
Discussion
The presence of ECS in HNSCC metastatic cervical nodes has a dramatic impact on patient outcomes and management. Currently, the ECS status in patients with metastatic HNSCC is determined postoperatively by histopathological evaluation of neck dissection specimens. The objective of this study was to assess the sensitivity and specificity of preoperative CT for ECS in cervical nodes with metastatic HNSCC. Sensitivity and specificity were determined by correlating radiological assessments of overall ECS status with histopathological findings. Radiological assessments of CT images were performed separately by two neuroradiologists. Our observers had around 80% specificity for overall assessment of ECS, with sensitivities around 60%. Consensus agreement was effective for improving accuracy, and the observers’ overall assessments were in agreement on 80% of the cases, 74 out of 93. During analysis of our observers’ impressions, it became apparent that in every case that an observer identified three or more ECS imaging criteria, they would also make an overall assessment of ECS positive. This occurred in about one third of all cases in our cohort, and of these cases in which three or more imaging criteria were identified about 86% had histopathological ECS. Correlation of the presence of three or more ECS imaging criteria with histopathological findings resulted in a higher specificity and PPV than overall observer ECS assessment.
Recent studies have examined the diagnostic accuracy of CT for ECS in patients with metastatic HNSCC, and are summarized in Table 6. Several of these studies demonstrate greater specificities than sensitivity for utilizing preoperative CT to identify ECS in HNSCC patients. Url et al. demonstrated specificities of 91% with sensitivities around 70%, by employing stringent imaging criteria including fat and soft tissue infiltration or infiltration of the sternocleidomastoid muscles, internal jugular vein or carotid artery.14 Our study employed these and additional criteria for qualifying nodal architecture, such as central necrosis, indistinct margins, and irregular nodal enhancement. Souter et al. found specificities of 91% and 90%, with sensitivities of 66% and 80% and concluded that preoperative radiological CT detection is applicable for informing patients, planning adjuvant treatment, and management guidance.15 King et al. also found high specificity with poor sensitivity by employing similar ECS imaging criteria.16 Prabhu et al. also demonstrated high specificity and low sensitivity, and found that CT performance improved directly with the histopathological severity of ECS.17 Similarly, our specificity and PPV improved when three or more ECS criteria were present. These results indicate that when compelling evidence exists on CT, such as gross invasion of adjacent structures, that these findings can be communicated to the patient care team as concerning for ECS, which may prove useful for preoperative management decisions. In contrast, an assessment of ECS by CT without compelling evidence, such as only a single ECS criterion finding, will result in higher false positive rates, which could negatively impact outcomes.
Table 6.
Summary table of studies evaluating CT detection of ECS in patients with HNSCC.
| Author | CT criteria | Sample size (ECS+) | Sn/Spa | PPV/NPV | Agreement (kappa) | Conclusion for preop CT detectability of ECS |
|---|---|---|---|---|---|---|
| King16 | IM, AI, IE | 51 (37) | CT MR 65/93 78/86 | –/– | 0.6, 0.5 | Comparable with MRI |
| Souter15 | IM, IE, PFC, thick-wall | 113 (N/A) | 66, 80/91, 90 | 85, 87/–,– | 0.67 | Radiologist experience improves accuracy |
| Url14 | AI | 49 (17) | 71, 76/91, 91 | –/– | 0.86 | High Sp |
| Chai18 | IM, AI, CI | 100 (63) | 49, 65/84, 54 | 84, 71/49, 48 | 0.37 | Not reliable |
| Prabhu17 | AI, IB, PFC | 432 (87) | 44/98 | 83/87 | – | High Sp; detectability rises with ECS grade |
| Randall20 | IM, CI, IE, CN | 40 (17) | For CN 91/50 | CN 59/88 | 0.49 CN 0.71 | CN on CT correlates with ECS in oral SCC |
| Aiken21 | AI, IB, PFC | 111 (29) | 68/88 for CN 66/93 | 66/89 | 0.86 | High Sp in oral HNSCC. CN is best criterion |
| Maxwell19 | IM, AI, PFC, CI | 65 (38) | 55, 47/70, 85 | 72, 82/53, 53 | 0.37 | Not reliable in p16+ |
| Carlton | IM, AI, IE, CN, matted | 93 (56) | 57,66/81,76 3+ criteria: 41, 54/89, 86 | 82,80/56,60 3+ criteria: 85, 86/50, 55 | 0.59 | Moderate Sp; 3+ criteria improves Sp |
CT: computed tomography; MRI: magnetic resonance imaging; ECS: extracapsular spread; HNSCC: squamous cell carcinoma of the head and neck; HP: histopathological; SCC: squamous cell carcinoma; IM: indistinct margins; AI: adjacent invasion; IB: irregular borders; CI: contour irregularity; IE: irregular/marginal enhancement; PFC: perinodal fat changes; CN: central necrosis; Sn: sensitivity; Sp: specificity.
In studies that reported results for individual reviewers, the results are included as results for observer 1, results for observer 2.
Compared to other studies included in Table 6, our observers had relatively lower sensitivity, particularly observer 1. One possible cause is that central necrosis was not as convincing to observer 1. There were four cases in which observer 1 identified central necrosis as the only ECS imaging criteria and made an overall false negative ECS assessment. In the same four cases observer 2 identified multiple ECS criteria including central necrosis and made an overall true positive ECS assessment. This disagreement exemplifies that between observers there are varying degrees of threshold for assessing nodal characteristics. Varying degrees of assessment threshold between our observers and between observers in other studies may contribute to our lower sensitivity. Also, our observers assessed for five ECS imaging criteria and offered an overall impression, while observers in other studies assessed for fewer criteria. When our observers detected only one or two imaging criteria, they simultaneously acknowledged that a greater proportion of ECS criteria were absent. A greater number of criteria being assessed allows for a greater proportion of absent findings to prompt a false negative overall assessment and lower sensitivity. Other studies assessed for fewer imaging criteria, which decreases the potential for a proportional influence.
Our sensitivity was also reduced by microscopic ECS. Prabhu et al. found that the sensitivity of CT detection of ECS was 18.8% when the extent of histopathological ECS 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.17 Figure 1(c) depicts a case in our cohort in which the microscopic extent of ECS was imperceptible to our observers, and a false negative assessment was made. It is possible that our study included a greater proportion of microscopic ECS compared to other studies, which would reduce the sensitivity and increased false negative rates. In consideration of false positives, Figure 1(d) depicts a case in our cohort in which multiple abnormal imaging findings were identified; however, histopathological assessment of ECS was negative. It is conceivable that histopathological assessment may not have detected ECS that was actually present. Another possibility is that the findings of ECS imaging criteria may be generated by processes other than ECS, and prompt a false positive overall ECS assessment. These processes could hypothetically include benign perinodal inflammation, vascular or lymphatic compression by mass effect, or central necrosis with preserved capsular integrity. In spite of potentially confounding processes, the high specificity demonstrated by our study and studies referenced in Table 6 indicate that false positives are rare when there is compelling findings to suggest ECS.
A study by Chai et al. used similar methods and criteria for assessing ECS status with preoperative CT.18 One of their observer’s results did not follow the sensitivity and specificity trends as in other studies, while the other observer’s results demonstrated high specificity and low sensitivity. Consequently, they found low interobserver agreement for overall ECS assessment and for imaging criteria, and their study concluded that CT assessment for ECS should not impact management. A subsequent study by Maxwell et al. assessed the accuracy of CT ECS assessment in a p16-positive HNSCC patient.19 One observer demonstrated sensitivity and specificity of CT ECS assessment comparable to the trends in other studies, while the observer had lower specificity and higher sensitivity. In both studies, differences in interobserver results can be accounted for by differences in observer thresholds for ECS assessment, which reciprocally affected PPV and the false positive rate.
In our study, interobserver agreement for overall ECS assessment was moderate, with a kappa value of 0.59. Prior studies also employed Cohen’s kappa for the measurement of interobserver agreement regarding ECS assessment using preoperative CT, as seen in Table 6. The variability in interobserver agreement between studies represents the current subjectivity in radiological ECS assessment. In an attempt to generate objectivity in ECS assessment, we evaluated the interobserver agreement for five traditional CT ECS imaging criteria. Our results demonstrated substantial agreement for central necrosis. A study by Randall et al. also found substantial interobserver agreement on central nodal necrosis, with a kappa of 0.72, and their interobserver agreement for overall ECS presence was also moderate, with a kappa of 0.49.20 Their results and ours reiterate that substantial interobserver agreement exists among neuroradiologists regarding CT assessment of central necrosis. Of note, both Randall et al. and Aiken et al. found that central necrosis was the best predictor of ECS among multiple other ECS imaging criteria.21 Aside from central necrosis, our interobserver agreement for other traditionally accepted CT features of ECS was at best moderate. Indistinct nodal margins and irregular nodal enhancement had the lowest interrater agreement and these criteria are not highly specific for ECS. The low interobserver agreement for these features further argues that these ECS criteria findings exclusively are not reliable for accurate reproducible ECS assessment.
There are multiple apparent limitations to our study. This was a retrospective study, which generates selection bias. In an attempt to minimize this bias, patients were selected on a chronological basis, and the list of images provided to the observers had been randomly assigned by date. Future prospective cohort studies are indicated to lessen this bias. In our study there was no attempt to correlate the exact nodes assessed by CT with nodes described in histopathological examination. However, eliminating this limitation would prove very difficult as during neck dissection tissue architecture and nodal arrangement is often altered and after histopathological processing the configuration of the nodes would be different than in situ. Another limitation in our study was that CT imaging was performed within 8 weeks prior to neck dissection. This time frame was chosen to minimize the potential for ECS to develop between imaging and operation, and to maximize sample size. It is possible that ECS could develop within 8 weeks from imaging to dissection. Future studies with shorter time frames from imaging to surgery would lessen this confounding element. Regarding the identification of ECS radiological criteria, there is not a defined standard for assessing the imaging criteria, which explains the variable interobserver agreements in our study as well as in others. Objective definitions for determining the presence of ECS imaging criteria would allow for increased interobserver agreement.
With regard to the prognostic value of ECS, recent studies have examined the prognostic value of ECS for patients with specifically oropharyngeal HNSCC. Sinha et al. found that ECS was not a prognostic indicator for a cohort of 152 patients with p16-positive oropharyngeal carcinoma, regardless of adjuvant therapy.22 Furthermore, Maxwell et al. found that in a cohort of 133 patients with oropharyngeal HNSCC, disease-specific survival was not associated with ECS status.23 Of note, in the same study, disease-specific survival was confirmed to be worse for patients with oral cavity HNSCC who had metastatic ECS in cervical lymph nodes. Based on the results of these studies, preoperative management considerations of CT assessment of ECS for patients with oropharyngeal HNSCC may not be warranted.
In all of the aforementioned studies, as well as ours, histopathological assessment of neck dissection specimens served as the standard. Brekel et al. demonstrated that interobserver agreement of histopathological ECS assessment was wide ranging, with kappas of 0.14–0.75.24 Given the histopathological assessment variability and HNSCC subgroup differences in ECS prognostic value, a study that examines the prognostic value of CT ECS assessment alone, and in correlation with histopathology, would further define the role of ESC assessment of preoperative CT in HNSCC management.
Conclusion
Intravenous contrast-enhanced CT imaging for the purpose of identifying ECS in positive cervical nodes for HNSCC has moderate specificity. The identification of three or more radiological criteria raises specificity and PPV, suggesting that CT performance is dependent on the severity of actual ECS. Interobserver agreement is moderate for overall ECS assessment and substantial for central necrosis. Interobserver agreement for indistinct nodal margins, infiltration into adjacent tissues, irregular nodal enhancement, and matted nodes is moderate at best, therefore these features should not be used individually as criteria for detecting ECS by CT.
Author contributions
JAC: project development, data collection, manuscript writing; AWM: data collection, manuscript writing; LBBauer: data collection; SMM: data collection; LJL: project development; HA: project development, data collection; AA: project development, data collection, manuscript writing.
Ethics
The authors declare that all human and animal studies have been approved by a University of Missouri institutional review board and have therefore been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. Informed consent was waived prior to inclusion in this study.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
References
- 1.Puri SK, Fan CY, Hanna E. Significance of extracapsular lymph node metastases in patients with head and neck squamous cell carcinoma. Curr Opin Otolaryngol Head Neck Surg 2003; 2: 119–123. [DOI] [PubMed] [Google Scholar]
- 2.Johnson JT, Barnes EL, Myers EN, et al. The extracapsular spread of tumors in cervical node metastasis. Arch Otolaryngol 1981; 107: 725–729. [DOI] [PubMed] [Google Scholar]
- 3.Liao CT, Huang SF, Chen IH, et al. Risk stratification of patients with oral cavity squamous cell carcinoma and contralateral neck recurrence following radical surgery. Ann Surg Oncol 2009; 16: 159–170. [DOI] [PubMed] [Google Scholar]
- 4.Liao CT, Lin CY, Fan KH, et al. Outcome analyses of unusual site neck recurrence in oral cavity cancer. Ann Surg Oncol 2013; 20: 257–266. [DOI] [PubMed] [Google Scholar]
- 5.Liao CT, Chang JT, Wang HM, et al. Preoperative [18F]fluorodeoxyglucose positron emission tomography standardized uptake value of neck lymph nodes predicts neck cancer control and survival rates in patients with oral cavity squamous cell carcinoma and pathologically positive lymph nodes. Int J Radiat Oncol Biol Phys 2009; 74: 1054–1061. [DOI] [PubMed] [Google Scholar]
- 6.Kuhnt T, Klockenbrink U, Knipping S, et al. Adjuvant low single dose cisplatin-based concurrent radiochemotherapy of oral cavity and oropharynx carcinoma: impact of extracapsular nodal spread on distant metastases. Strahlenther Onkol 2011; 187: 292–299. [DOI] [PubMed] [Google Scholar]
- 7.Coatesworth AP, MacLennan K. Squamous cell carcinoma of the upper aerodigestive tract: the prevalence of microscopic extracapsular spread and soft tissue deposits in the clinically N0 neck. Head and Neck 2002; 3: 229–254. [DOI] [PubMed] [Google Scholar]
- 8.Lewis JS, Jr, Carpenter DH, Thorstad WL, et al. Extracapsular extension is a poor predictor of disease recurrence in surgically treated oropharyngeal squamous cell carcinoma. Modern Pathol 2011; 11: 1413–1420. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Greenberg JS, Fowler R, Gomez J, et al. Extent of extracapsular spread: a critical prognosticator in oral tongue cancer. Cancer 2003; 6: 1464–1470. [DOI] [PubMed] [Google Scholar]
- 10.National Comprehensive Cancer Network. Head and Neck Cancers (version 2.2016). http://www.nccn.org/professionals/physician_gls/PDF/head-and-neck.pdf. (accessed 21 December 2016).
- 11.Curtin HD, Ishwaran H, Mancuso AA, et al. Comparison of CT and MR imaging in staging of neck metastases. Radiology 1998; 207: 123–130. [DOI] [PubMed] [Google Scholar]
- 12.Bernier J, Cooper JS, Pajak TF, et al. Defining risk levels in locally advanced head and neck cancers: a comparative analysis of concurrent postoperative radiation plus chemotherapy trials of the EORTC (#22931) and RTOG (#9501). Head Neck 2005; 27: 843–850. [DOI] [PubMed] [Google Scholar]
- 13.Steinkamp HJ, Beck A, Werk M, et al. Extracapsular spread of cervical lymph node metastases: diagnostic value of magnetic resonance imaging. Rofo 2002; 1: 50–55. [DOI] [PubMed] [Google Scholar]
- 14.Url C, Schartinger V, Riechelmann H, et al. Radiological detection of extracapsular spread in head and neck squamous cell carcinoma (HNSCC) cervical metastases. Eur J Radiol 2013; 82: 1783–1787. [DOI] [PubMed] [Google Scholar]
- 15.Souter MA, Allison RS, Clarkson JH, et al. Sensitivity and specificity of computed tomography for detection of extranodal spread from metastatic head and neck squamous cell carcinoma. J Laryngol Otol 2009; 123: 778–782. [DOI] [PubMed] [Google Scholar]
- 16.King AD, Tse GM, Yuen EH, et al. Comparison of CT and MR imaging for the detection of extranodal neoplastic spread in metastatic neck nodes. Eur J Radiol 2004; 3: 264–270. [DOI] [PubMed] [Google Scholar]
- 17.Prabhu RS, Magliocca KR, Hanasoge S, et al. Accuracy of computed tomography for predicting pathologic nodal extracapsular extension in patients with head-and-neck cancer undergoing initial surgical resection. Int J Radiat Oncol Biol Phys 2014; 88: 122–129. [DOI] [PubMed] [Google Scholar]
- 18.Chai RL, Rath TJ, Johnson JT, et al. Accuracy of computed tomography in the prediction of extracapsular spread of lymph node metastases in squamous cell carcinoma of the head and neck. JAMA Otolaryngol Head Neck Surg 2013; 139: 1187–1194. [DOI] [PubMed] [Google Scholar]
- 19.Maxwell JH, Rath TJ, Byrd JK, et al. Accuracy of computed tomography to predict extracapsular spread in p16-positive squamous cell carcinoma. The Laryngoscope 2015; 125: 1613–1618. [DOI] [PubMed] [Google Scholar]
- 20.Randall D, Lysack J, Hudon M, et al. Diagnostic utility of central node necrosis in predicting extracapsular spread among oral cavity squamous cell carcinoma. Head Neck 2015; 37: 92–96. [DOI] [PubMed] [Google Scholar]
- 21.Aiken AH, Poliashenko S, Beitler JJ, et al. Accuracy of preoperative imaging in detecting nodal extracapsular spread in oral cavity squamous cell carcinoma. Am J Neuroradiol 2015; 36: 1776–1781. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Sinha P, Lewis JS, Piccirillo JF, et al. Extracapsular spread and adjuvant therapy in human papillomavirus-related, p16-positive oropharyngeal carcinoma. Cancer 2012; 118: 3519–3530. DOI: 10.1002/cncr.26671. [DOI] [PubMed] [Google Scholar]
- 23.Maxwell JH, Ferris RL, Gooding W, et al. Extracapsular spread in head and neck carcinoma: impact of site and human papillomavirus status. Cancer 2013; 119: 3302–3308. [DOI] [PubMed] [Google Scholar]
- 24.Van den Brekel MW, Lodder WL, Stel HV, et al. Observer variation in the histopathologic assessment of extranodal tumor spread in lymph node metastases in the neck. Head Neck 2012; 34: 840–845. [DOI] [PubMed] [Google Scholar]