Diagnostic accuracy of computed tomography alone versus computed tomography combined with ultrasound for diagnosing lymph node metastasis in thyroid cancer Authors: Hussain A. Alharbi, Laila Z. Alzamil, Sumaiya H. Muathen, Hadad H. Alkaff, Omar AbuSuliman, Faisal A. Saati, Ameen Z. Alherabi Published: 2026 ABSTRACT ObjectivesTo evaluate and compare the diagnostic accuracy of computed tomography alone versus computed tomography combined with ultrasound for detecting metastatic cervical lymph nodes in patients with thyroid cancer using the Standards for Reporting Diagnostic Accuracy Studies 2015 guidelines.MethodsThis retrospective diagnostic accuracy study included 134 patients with thyroid cancer who underwent preoperative computed tomography or ultrasound scans for lymph node metastases at King Abdullah Medical City Hospital, Makkah, KSA, between 2009 and 2020. Histopathological examination was the reference standard. Sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy were assessed for computed tomography alone versus computed tomography combined with ultrasound.ResultsThe mean age of the participants was 46 years, and over two-thirds were females. Papillary thyroid carcinoma was the most common diagnosis, followed by follicular and medullary thyroid carcinomas. Lymphovascular invasion was observed in 32% of cases. The sensitivity, specificity, positive predictive value, negative predictive value, accuracy, and area under the curve were 71.05%, 91.67%, 77.14%, 88.89%, 85.82%, and 0.814, respectively, for computed tomography + ultrasound; and 28.95%, 98.96%, 91.67%, 77.87%, 79.10%, and 0.64 for central cervical nodes (zone VI), and 73.68%, 90.63%, 75.68%, 89.69%, 85.82%, and 0.822 for peripheral cervical nodes.ConclusionsComputed tomography combined with ultrasound exhibited acceptable diagnostic performance in detecting metastatic cervical lymph nodes in thyroid cancer, but the sensitivity of computed tomography alone was lower for central cervical nodes. Ultrasound-only analysis was not performed, so these findings should be interpreted as supportive of a complementary, multimodal imaging approach rather than as evidence that computed tomography combined with ultrasound is superior to ultrasound alone. Introduction Cervical lymph node metastasis (CLNM) is relatively common in thyroid cancer, particularly in papillary thyroid carcinoma (PTC), which accounts for ∼90% of all cases.1 The occurrence of CLNM during thyroid cancer progression varies widely depending on the tumor type and stage. Studies indicate that ∼20–50% of patients with PTC present with CLNM at the time of diagnosis.2, 3, 4 Conversely, follicular thyroid carcinoma is characterized by lower rates of cervical lymph node involvement, typically ranging from 10% to 20%.5 Therefore, accurately assessing CLNM and determining the appropriate extent of neck dissection is crucial for optimizing clinical outcomes and refining the surgical approach in thyroid cancer.6, 7, 8 Therapeutic intervention for thyroid cancer involves central or lateral compartment neck dissection when lymph nodes show clinical involvement or high-risk tumor characteristics are detected through imaging or clinical assessment. The American Thyroid Association (ATA) guidelines recommend that prophylactic cervical lymph node dissection should not be conducted routinely in patients with thyroid cancer unless there is definite clinical evidence of metastasis, confirmed through imaging studies or pathological examination.9 According to the ATA guidelines, preoperative staging of CLNM should prioritize the use of high-resolution ultrasound (US) as the primary tool to assess cervical lymph nodes in patients with thyroid cancer. This method effectively identifies suspicious nodes based on characteristics such as size, shape, echogenicity, and vascular patterns, and should be applied across all cervical lymph node compartments, including the central and lateral neck areas. If a suspicious node is identified, US-guided needle aspiration biopsy can be performed immediately during the same procedure.9 According to previous studies, the diagnostic accuracy of US for detecting cervical nodal metastasis in thyroid cancer varies, with a sensitivity of 56% and specificity of ∼93%.10 US has advantages such as the absence of radiation exposure, making it safe for repeated use, and providing real-time imaging. However, it is operator-dependent and has limitations, including a restricted field of view and challenges when assessing deeper or posterior nodes, particularly in patients with large body habitus or significant neck swelling. Conversely, computed tomography (CT) scans can address these limitations by providing detailed imaging and reducing the dependence on operator skills, enabling comprehensive evaluation of lymph nodes. However, CT scans have disadvantages, such as potential risks associated with repeated radiation exposure, nephrotoxicity from contrast agents, and lower accessibility compared with US.11 Several studies have examined the diagnostic accuracy of CT scans combined with US at detecting CLNM in patients with thyroid cancer.12, 13, 14 However, to the best of our knowledge, no studies have been conducted to assess the accuracy of detecting CLNM using CT and US in KSA, following the Standards for Reporting Diagnostic Accuracy Studies (STARD) guidelines.15 Therefore, in the present study, we aimed to evaluate and compare the diagnostic accuracy of CT alone and CT combined with US for assessing metastatic cervical lymph nodes in patients with thyroid cancer in KSA using the STARD guidelines for reporting diagnostic accuracy tests. This retrospective study aimed to evaluate the diagnostic accuracy, including the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV), of CT alone and CT combined with US (CT + US) for detecting metastatic cervical lymph nodes in patients with thyroid cancer. Materials and Methods Study design and reporting We conducted a retrospective diagnostic test accuracy study at King Abdullah Medical City Hospital, Makkah, KSA. The study was reported in accordance with the STARD 2015 guidelines.15 A completed STARD 2015 checklist is provided as Supplementary Material to document the completeness of reporting. Prospective recruitment, prespecified sample size calculation, and standardized blinding procedures could not be implemented due to the retrospective design and reliance on routinely collected clinical records. Study setting and patient selection Data were extracted from the medical records of King Abdullah Medical City Hospital, Makkah, KSA, which is a tertiary public referral center, and specifically from the Head and Neck and Skull Base Center. Eligible patients were identified retrospectively from their medical records at the Head and Neck and Skull Base Center of King Abdullah Medical City Hospital between 2009 and 2020. Patients were included if they had a histopathologically confirmed diagnosis of thyroid cancer and underwent preoperative imaging assessment for CLNM using CT, US, or combined CT and US. Patients were excluded if they had incomplete imaging data, missing histopathological confirmation, or inadequate clinical records. Index tests The index tests were preoperative CT alone and CT combined with US for detecting CLNM. CT was used for anatomical assessment of cervical lymph nodes, including central and peripheral compartments. US was used to assess suspicious lymph node features, including morphology and other sonographic characteristics. Imaging findings were classified as positive or suspicious for CLNM or negative based on the available radiology reports. The imaging protocols, reporting criteria, and operator experience could not be fully standardized across all cases because this was a retrospective study. In addition, blinding of radiologists to clinical or pathological information could not be reliably confirmed from the available records. Reference standard Histopathological examination of cervical lymph nodes obtained through biopsy or surgical excision was used as the reference standard for confirming CLNM. Imaging findings were compared with histopathological results to determine true-positive, false-positive, true-negative, and false-negative cases. Data collection The extracted variables included age, sex, nationality, city of residence, incidental onset, duration of illness, compression symptoms, previous thyroid surgery, family history of thyroid cancer, histopathological diagnosis, lymphovascular invasion, tumor focality, tumor size, tumor stage, nodal stage, distant metastasis status, Metastasis, Age, Completeness of resection, Invasion, and Size (MACIS) score, and Age, Metastases, Extent, and Size (AMES) prognostic classification. Diagnostic data were extracted for CT alone and combined CT + US where available. US-only diagnostic data were not consistently available in the medical records, so a separate US-only diagnostic accuracy analysis could not be performed. Diagnostic accuracy measures Diagnostic performance was evaluated based on the sensitivity, specificity, PPV, NPV, overall accuracy, and area under the receiver operating characteristic curve (AUC). These measures were calculated using histopathology as the reference standard. The diagnostic performance was assessed for:1.Combined CT + US for detecting suspicious cervical lymph nodes.2.CT alone for central cervical lymph nodes, including zone VI.3.CT alone for peripheral cervical lymph nodes, including levels I–V and VII. 1.Combined CT + US for detecting suspicious cervical lymph nodes. Combined CT + US for detecting suspicious cervical lymph nodes. 2.CT alone for central cervical lymph nodes, including zone VI. CT alone for central cervical lymph nodes, including zone VI. 3.CT alone for peripheral cervical lymph nodes, including levels I–V and VII. CT alone for peripheral cervical lymph nodes, including levels I–V and VII. Statistical analysis All statistical analyses were performed using SPSS version 26 (SPSS Inc., Chicago, IL, USA). Descriptive statistics were used to summarize baseline demographic and clinicopathological characteristics. Categorical variables were presented as frequencies and percentages. Diagnostic accuracy parameters were calculated using 2 × 2 contingency tables to compare imaging findings with histopathological results. True-positive, false-positive, true-negative, and false-negative results were used to calculate the sensitivity, specificity, PPV, NPV, and overall accuracy. Receiver operating characteristic (ROC) analysis was performed, and the AUC was used to summarize overall diagnostic performance. Point estimates were made for diagnostic accuracy measures. It was not possible to calculate 95% confidence intervals for the sensitivity, specificity, PPV, NPV, accuracy, and AUC because the original statistical output and available retrospective data set did not contain the information required to derive reliable interval estimates, and thus only point estimates could be reported. Additional adjusted analyses or subgroup analyses could not be reliably performed due to the retrospective nature of the data set and the limited availability of complete US-only and covariate-level diagnostic data. Results Baseline characteristics of participants In this retrospective diagnostic test accuracy study, 134 patients were enrolled with various types of thyroid cancer. The majority were aged 46 years or older (44.1%; 59/134), followed by those aged 20–39 years (38.8%; 52/134). Participants aged 40–45 years accounted for 13.4% (18/134) and 3.7% (5/134) were aged 19 years or younger. Females comprised over two-thirds of the study participants (91/134). In addition, 73.9% (99/134) were Saudi whereas 26.1% were non-Saudis. The primary residential cities for most of the participants were Jeddah 68 (51%) and Makkah 38 (28.4%). Approximately 54% of the participants had incidental onset thyroid cancer (72/134), and only 13% (17/134) experienced compression symptoms. PTC was the most common diagnosis, accounting for 78% (105/134), followed by follicular (11.1%; 15/134) and medullary thyroid carcinoma (4.4%; 6/134). Lymphovascular invasion was observed in 32% (43/134), whereas no evidence of invasion was found in 68% (91/134). Most tumors were unifocal, accounting for 76 participants, whereas multifocal tumors were observed in 58 participants. The distribution of tumor sizes in ascending order was as follows: <1 cm = 17% (23/134); 1–2 cm = 28% (38/134); 2–4 cm = 30% (40/134); and >4 cm = 25% (33/134). The pathological staging of tumors based on size followed the order of: T1a = 17% (23/134); T1b = 28.5% (38/134); T2 = 27% (36/134); T3 = 25.4% (34/134); T4a = 1.4% (2/134); and T4b = 0.7% (1/134). Nodal metastasis was identified in 28% (38/134) of the study participants. Ten cases of nodal metastasis were identified in zone six (zone VI) of the cervical lymph nodes, representing the central compartment. In addition, 28 cases were identified in the peripheral compartment of the cervical lymph nodes. The detailed characteristics of the participants are presented in Table 1.Table 1Study characteristics and baselines of the participants.Table 1VariablesN (%)Age≤195 (3.7%)20–3952 (38.8)40–4518 (13.4%)≥ 4659 (44.1%)GenderMale43 (32%)Female91 (68%)NationalitySaudi99 (73.9%)Non-Saudi35 (26.1%)City of residenceJeddah68 (51%)Makkah38 (28.4%)Taif9 (6.7%)Albaha8 (6%)Jezan3 (2.2%)Alqonfetha3 (2.2%)Others5 (3.5%)Incidental onsetYes72 (53.7%)No62 (46.3%)Compression symptomsYes17 (13%)No117 (87%)Previous outside thyroid surgeryYes27 (20%)No107 (80%)Family history of carcinomaYes1 (0.7%)No133 (99.3%)Histopathology of the lesionPapillary105 (78.7%)Follicular15 (11.1%)Hurthle cell6 (4.4%)Medullary6 (4.4%)Anaplastic2 (1.4%)Lymphovascular invasionYes43 (32%)No91 (68%)Tumor focalityUnifocal76 (56.7%)MultiFocal58 (43.3%)Tumor pathological size (cm)<123 (17%)1 to 238 (28%)2 to 440 (30%)>433 (25%)Tumor (T) staging (pathological size)T1a23 (17%)T1b38 (28.5%)T236 (27%)T334 (25.4%)T4a2 (1.4%)T4b1 (0.7%)Nodal "N" stagingN096 (72%)N1a10 (7%)N1b28 (21%)Metastasis "M" stagingM0127 (95%)M17 (5%)Tumor stageI97 (72.7%)II7 (5%)III16 (12%)IVa10 (7.4%)IVb1 (0.7%)IVc3 (2.2%)MACIS score<6 (low risk)112 (83.5%)6–7 (intermediate risk)10 (7.5%)>7 (high risk)12 (9%)AMESLow risk96 (61.7%)Intermediate risk1 (0.7%)High risk37 (27.6%)Abbreviations: AMES = Age, Metastases, Extent and Size; MACIS = distant Metastasis, patient Age, Completeness of resection, local Invasion, and tumor Size; N0 = (No regional lymph node metastasis); N1a = (Mets to level VI (pre tracheal, paratracheal, and pre-laryngeal/Delphian lymph nodes); N1b = (Mets to unilateral, bilateral, or contralateral cervical (levels I, II, III, IV, or V) or superior mediastinal lymph nodes (level VII); M0 = (No distant metastasis is found); M1 = (Distant metastasis is present). Table 1Study characteristics and baselines of the participants.Table 1VariablesN (%)Age≤195 (3.7%)20–3952 (38.8)40–4518 (13.4%)≥ 4659 (44.1%)GenderMale43 (32%)Female91 (68%)NationalitySaudi99 (73.9%)Non-Saudi35 (26.1%)City of residenceJeddah68 (51%)Makkah38 (28.4%)Taif9 (6.7%)Albaha8 (6%)Jezan3 (2.2%)Alqonfetha3 (2.2%)Others5 (3.5%)Incidental onsetYes72 (53.7%)No62 (46.3%)Compression symptomsYes17 (13%)No117 (87%)Previous outside thyroid surgeryYes27 (20%)No107 (80%)Family history of carcinomaYes1 (0.7%)No133 (99.3%)Histopathology of the lesionPapillary105 (78.7%)Follicular15 (11.1%)Hurthle cell6 (4.4%)Medullary6 (4.4%)Anaplastic2 (1.4%)Lymphovascular invasionYes43 (32%)No91 (68%)Tumor focalityUnifocal76 (56.7%)MultiFocal58 (43.3%)Tumor pathological size (cm)<123 (17%)1 to 238 (28%)2 to 440 (30%)>433 (25%)Tumor (T) staging (pathological size)T1a23 (17%)T1b38 (28.5%)T236 (27%)T334 (25.4%)T4a2 (1.4%)T4b1 (0.7%)Nodal "N" stagingN096 (72%)N1a10 (7%)N1b28 (21%)Metastasis "M" stagingM0127 (95%)M17 (5%)Tumor stageI97 (72.7%)II7 (5%)III16 (12%)IVa10 (7.4%)IVb1 (0.7%)IVc3 (2.2%)MACIS score<6 (low risk)112 (83.5%)6–7 (intermediate risk)10 (7.5%)>7 (high risk)12 (9%)AMESLow risk96 (61.7%)Intermediate risk1 (0.7%)High risk37 (27.6%)Abbreviations: AMES = Age, Metastases, Extent and Size; MACIS = distant Metastasis, patient Age, Completeness of resection, local Invasion, and tumor Size; N0 = (No regional lymph node metastasis); N1a = (Mets to level VI (pre tracheal, paratracheal, and pre-laryngeal/Delphian lymph nodes); N1b = (Mets to unilateral, bilateral, or contralateral cervical (levels I, II, III, IV, or V) or superior mediastinal lymph nodes (level VII); M0 = (No distant metastasis is found); M1 = (Distant metastasis is present). Study characteristics and baselines of the participants. Abbreviations: AMES = Age, Metastases, Extent and Size; MACIS = distant Metastasis, patient Age, Completeness of resection, local Invasion, and tumor Size; N0 = (No regional lymph node metastasis); N1a = (Mets to level VI (pre tracheal, paratracheal, and pre-laryngeal/Delphian lymph nodes); N1b = (Mets to unilateral, bilateral, or contralateral cervical (levels I, II, III, IV, or V) or superior mediastinal lymph nodes (level VII); M0 = (No distant metastasis is found); M1 = (Distant metastasis is present). Diagnostic test accuracy The diagnostic accuracies of imaging modalities for detecting lymph node metastasis were evaluated using the sensitivity, specificity, PPV, and NPV. The positive or suspicious lymph nodes identified by CT scan and US resulted in a sensitivity of 71.05%, specificity of 91.67%, PPV of 77.14%, NPV of 88.89%, accuracy of 85.82%, and AUC of 0.814. When focusing on central cervical nodes (zone VI) detected by CT scan, the sensitivity for this region was markedly lower at 28.95%, with a specificity of 98.96%, PPV of 91.67%, NPV of 77.87%, accuracy of 79.10%, and AUC of 0.64. For peripheral cervical nodes detected by CT scan, the results indicated a sensitivity of 73.68%, specificity of 90.63%, PPV of 75.68%, NPV of 89.69%, accuracy of 85.82%, and AUC of 0.822 (Table 2 and Figure 1).Table 2Diagnostic performance of CT and CT + US.Table 2ImagingTPTNFPFNSensitivity (%)Specificity (%)PPV (%)NPV (%)Accuracy (%)AUCPositive/Suspicious lymph nodes (by CT scan + US)278881171.0591.6777.1488.8985.820.814Nodes (region VI–central) (by CT scan)119512728.9598.9691.6777.8779.100.640Nodes (Regions I–V, or VII – other) (by CT scan)288791073.6890.6375.6889.6985.820.822Abbreviations: TP = true positive; TN = true negative; FP = false-positive; FN = false negative; PPV = positive predictive value; NPV = negative predictive value; AUC = area under the curve.Figure 1ROC Curve. Table 2Diagnostic performance of CT and CT + US.Table 2ImagingTPTNFPFNSensitivity (%)Specificity (%)PPV (%)NPV (%)Accuracy (%)AUCPositive/Suspicious lymph nodes (by CT scan + US)278881171.0591.6777.1488.8985.820.814Nodes (region VI–central) (by CT scan)119512728.9598.9691.6777.8779.100.640Nodes (Regions I–V, or VII – other) (by CT scan)288791073.6890.6375.6889.6985.820.822Abbreviations: TP = true positive; TN = true negative; FP = false-positive; FN = false negative; PPV = positive predictive value; NPV = negative predictive value; AUC = area under the curve. Diagnostic performance of CT and CT + US. Abbreviations: TP = true positive; TN = true negative; FP = false-positive; FN = false negative; PPV = positive predictive value; NPV = negative predictive value; AUC = area under the curve. ROC Curve. Discussion The present study to assess the diagnostic test accuracy when identifying lymph node metastases in various thyroid carcinomas utilized data from 134 patients, and approximately two-thirds were females. Our findings indicate that the diagnostic parameters obtained for CT combined with US when detecting suspicious lymph node metastases (LNMs) across the entire cervical region were comparable to those using CT alone in peripheral cervical lymph nodes. CT alone exhibited reasonable sensitivity and specificity for detecting LNMs in the peripheral cervical regions but CT + US provided a complementary view, making it preferable for more comprehensive assessment. However, using CT alone for detecting suspicious LNMs in the central cervical region exhibited significantly lower sensitivity and higher specificity compared with CT + US. Given the low sensitivity of CT alone, relying exclusively on it could result in missed diagnoses. Hence, a combined approach might be necessary to thoroughly evaluate the central region of cervical lymph nodes. Previously, Alabousi et al. conducted a large meta-analysis of 47 diagnostic test accuracy studies that enrolled 31,942 patients with thyroid cancer.16 Both CT and US alone produced relatively lower sensitivity and specificity parameters for the diagnosis of central cervical LNMs, but they found that CT scanning was associated with significantly higher sensitivity compared with US (39% vs 28%). However, the specificity of CT scanning for the central group of cervical LNMs was lower than that using US (87% vs 95%, respectively).16 Similarly, Bongers et al.17 evaluated 111 patients with clinically low-risk differentiated thyroid cancer and they showed that CT findings impacted surgical management in 25 cases (22.5%). In particular, 16 cases benefited from the detection and removal of significant lymph node disease that was not identified by US. Thus, for patients with clinically low-risk differentiated thyroid cancer, the addition of CT to preoperative staging can significantly impact surgical decisions by detecting nodal metastasis, which might be missed by US alone.17 Similarly, Lesnik et al. found that both CT and US exhibited low sensitivity and specificity when scanning the central compartment of cervical lymph nodes.18 Our study did not assess the diagnostic parameters with US alone for detecting cervical LNMs in thyroid carcinoma. However, a previous meta-analysis by Alabousi determined the sensitivity and specificity for US in diagnosing both central (28% and 95%, respectively) and peripheral (73% and 89%, respectively) cervical LNMs.16 In particular, they found no significant difference between using CT alone and US alone as diagnostic parameters for identifying cervical LNMs in terms of their sensitivity (77% vs. 73%) or specificity (88% vs. 89%).16 Conversely, we found that combining CT + US yielded a sensitivity of 71.05% and specificity of 91.67% for detecting cervical LNMs in patients with various thyroid carcinomas. Lu and Chen compared the diagnostic parameters for US, CT, and combined CT + US in the detection of cervical LNMs in patients with PTC.19 They found that combined CT + US exhibited greater sensitivity, specificity, and accuracy (99.44%, 71.43%, and 98.38%, respectively) compared with either US scanning (75.28%, 26.31%, and 75.13%, respectively) or CT scanning (78.09%, 26.67%, and 77.29%, respectively).19 In addition, CT scanning produced numerically superior but comparable diagnostic parameters compared with US scanning.19 Moreover, Wang et al. conducted a meta-analysis by pooling 11 individual studies that included a total of 6461 patients.20 They reported that combining US + CT had a sensitivity of 89% and specificity of 79% in diagnosing cervical LNMs in patients with various thyroid cancers. They also showed that CT + US exhibited relatively greater sensitivity and lower specificity than either US or CT alone. Considering that US alone is not sufficient for tumor staging and overcoming anatomical limitations regarding deep structures, they recommended using CT + US in the preoperative examination of various types of thyroid cancers.20 Evidence from previous retrospective studies,21 prospective studies,18,22,23 and meta-analyses24,25 also supports the superiority of CT + US over either CT or US alone in terms of the diagnostic parameters. US remains the primary imaging modality and current gold standard for the preoperative evaluation of CLNM in thyroid cancer. US is highly suitable for evaluating superficial structures due to its high-resolution imaging and ability to differentiate between cystic and solid nodules. It is also non-invasive, less expensive, and prevents exposure to ionizing radiation.25 However, the quality and accuracy of US depend greatly on the operator's expertise, making it a highly subjective procedure.26 In addition, US has limitations in identifying metastatic lymph nodes smaller than 1 cm or those located near bony structures, such as the mandible.22 By contrast, CT scans provide a more comprehensive view of both superficial and deep anatomical structures, yielding in-depth visualizations of tissues behind bones and muscles. In addition, CT can identify deeper lymph nodes, including retrosternal, retropharyngeal, and mediastinal lymph nodes. A CT scan is also less operator-dependent, resulting in more consistent diagnostic outcomes, which is particularly advantageous for identifying LNMs in deeper regions, such as the central neck compartment, where US may be less accurate. However, CT is more expensive, exposes patients to higher radiation levels, and generally takes longer than US, especially when contrast administration and image processing are considered.18,21,25,27 Overall, US remains an important first-line imaging modality for preoperative cervical lymph node assessment, particularly for superficial lymph nodes, because of its safety, accessibility, and the absence of radiation exposure. CT may provide additional anatomical information, especially for deeper or less accessible nodal regions. In the present study, CT alone exhibited reasonable diagnostic performance for peripheral cervical lymph nodes but low sensitivity for central cervical nodes, suggesting that CT alone may miss central compartment metastases. Therefore, the combined use of CT and US may provide complementary diagnostic information, although we cannot conclude that CT + US is superior to US alone because US-only diagnostic accuracy data were not available in our data set. Thus, these findings should be interpreted with caution and viewed as supportive of a multimodal imaging approach rather than as evidence for changing diagnostic protocols. Future prospective studies are needed with standardized CT and US protocols, separate US-only analysis, and adjustment for potential confounders to confirm the optimal imaging strategy. We conducted the first diagnostic study of this type in the Saudi population, providing valuable insights into the diagnostic parameters for combined US + CT or CT alone in identifying cervical LNMs in patients with various thyroid cancers. Our study included 134 patients, which is similar to the average sample size for studies and populations of this nature.16 However, our study had several limitations. The data were retrospective and extracted from patient records over time. We were unable to subgroup the combined CT and US data into either the central or peripheral cervical lymph node zones. In addition, selection bias cannot be excluded because this was a retrospective single-center study from a tertiary referral hospital. The inclusion of only patients with available imaging and histopathological confirmation records may limit the generalizability of the findings. Furthermore, the lack of external validation and potential confounding factors limits the robustness of the findings. Multi-center studies are recommended to address these limitations and enhance the applicability and reliability of the results. In particular, US scanning is operator-dependent, and the experience of the radiologist significantly affects the identification of cervical LNMs.26, 27, 28, 29, 30 However, our study did not account for variability in ultrasonographer expertise during the analysis, which could have introduced additional bias. This study was reported according to the STARD 2015 framework, but several STARD items could not be fully addressed because of the retrospective design. Patient recruitment was based on available clinical records, and thus a prespecified sample size calculation was not performed, and blinding between interpretation of the index tests and reference standard could not be reliably confirmed. These limitations may have increased the risk of bias and should be considered when interpreting the diagnostic accuracy estimates. A major limitation of this study is the absence of separate US-only diagnostic accuracy analysis. US was used as part of the preoperative evaluation in some patients, but US-only data were not consistently available in a form that allowed reliable calculations of the sensitivity, specificity, PPV, NPV, accuracy, or AUC. The absence of US-only data was clearly a reporting limitation but it also fundamentally restricts the ability to make direct comparisons between imaging modalities, particularly preventing any determination of whether combined CT + US provides a diagnostic advantage over US alone. Therefore, our findings should not be interpreted as a direct comparison between CT alone, US alone, and combined CT + US. Future prospective studies are needed with standardized imaging protocols to directly compare these modalities. Another limitation is that confidence intervals were not available for all diagnostic accuracy estimates. We could not derive 95% confidence intervals from the available retrospective data set, so the precision and statistical stability of the reported point estimates of sensitivity, specificity, PPV, NPV, accuracy, and AUC could not be assessed, and thus these values should be regarded as descriptive rather than definitive. This should be considered when interpreting the stability and clinical applicability of the reported diagnostic performance measures. Potential confounding factors were not adjusted for in the present analysis. Variables such as tumor size, nodal compartment, histological subtype, lymphovascular invasion, imaging quality, and radiologist or ultrasonographer experience may have influenced diagnostic performance. Multivariable adjustment was not feasible because of the retrospective design, limited sample size, and incomplete covariate-level diagnostic data. Therefore, the reported diagnostic accuracy estimates should be interpreted with caution. Conclusion In the present study, we found that combined CT + US exhibited acceptable diagnostic performance at detecting CLNM in thyroid cancer, whereas the performance of CT alone differed according to the nodal compartment. CT alone exhibited reasonable diagnostic accuracy for peripheral cervical lymph nodes but low sensitivity for central cervical lymph nodes, suggesting that central nodal disease may be missed when relying on CT alone. However, US-only diagnostic accuracy data were not available, so these findings should not be interpreted as direct comparisons of CT alone, US alone, and combined CT + US. Therefore, combined CT + US may provide complementary diagnostic value in preoperative assessment, but the results should be interpreted with caution. Further prospective studies are needed with standardized CT and US protocols to validate these findings, including separate US-only analysis. Ethical approval This study was approved by the Institutional Review Board (IRB) of King Abdullah Medical City and registered at the National BioMedical Ethics Committee under IRB number 14–084. The requirement for informed consent was waived due to the retrospective nature of the study and the use of anonymized clinical data. Authors contributions HAA contributed to study conception, data collection, and manuscript drafting. LZA and SHM contributed to data analysis and interpretation. HHA and OA contributed to study design and clinical data acquisition. FAS contributed to data interpretation and manuscript revision. AZA supervised the study and critically revised the manuscript. All authors have critically reviewed and approved the final draft and are responsible for the content and similarity index of the manuscript. Source of funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Conflict of interest There are no conflicts of interest to declare.