The value of nomogram analysis in predicting pulmonary metastasis in hepatic alveolar echinococcosis

Categories: Disease & Virus

May 14, 2025

General information

This study was approved by the Ethics Committee of the Clinical Medical College of Qinghai University (Approval No: P-SL-2023-229). Due to the retrospective nature of the study, the requirement for informed consent was waived by the Ethics Committee of the Clinical Medical College of Qinghai University. The study protocol followed the principles of the Declaration of Helsinki, and all procedures complied with the national and institutional ethical standards.

Retrospectively collected imaging features and clinical indicators of patients first diagnosed with HAE at the Qinghai University Affiliated Hospital from 2015 to 2022. Inclusion criteria: (1) Pathologically confirmed as HAE after surgery or meeting the diagnostic criteria in the “Expert Consensus on Imaging Diagnosis of Hepatic Echinococcosis”¹⁷; (2) Complete clinical data of the patients; (3) All patients underwent dynamic contrast-enhanced CT scanning of the abdominal cavity in three phases and chest CT examination. Exclusion criteria: (1) Presence of other tumorous diseases in the liver. (2) Inability to evaluate due to severe image artifacts, as shown in Fig. 1. Two hundred ninety-seven patients with hepatic alveolar echinococcosis were included, comprising 138 males and 159 females, with an average age of 39.23 ± 14.03 years. Using the “Random Number Generator” in SPSS software, patients were randomly allocated to the training set and test set in a 7:3 ratio.Subsequently, the training set data was used for model construction, and the test set data was used to verify the model’s effectiveness.

Examination methods

CT examination

All patients underwent abdominal three-phase dynamic contrast-enhanced CT scanning and chest CT plain scan with enhancement using a 128-slice spiral CT scanner (Discovery CT 750 HD; GE Medical Systems, Milwaukee, Wis). The abdominal scanning range extended from the rib cage to the iliac bones, with patients lying supine. The scanning parameters were as follows: tube voltage of 80-120KV, tube current of 200 mA, slice thickness of 5 mm, and rotation time of 0.5 s per revolution. First, a plain scan was performed, followed by a contrast-enhanced scan. The contrast agent used was ioversol injection (350 mg L/ml), administered through the median cubital vein at a 4–5 ml/s rate. Threshold tracking was employed for scanning, and once the threshold of 120HU was reached, the scan was automatically triggered. With a minimum delay of 2 s after reaching the threshold, the first phase (aortic phase) was scanned. Subsequently, after a delay of 25 s, the second phase (portal venous phase) was scanned, followed by a delay of 40 s for the third phase (delayed phase) scan. The scanning range for the chest encompassed the thoracic inlet to the base of the pulmonary region, with the patient positioned in a supine layout. The scanning parameters were set as follows: a tube voltage of 120 KV, a tube current of 10 mA, a slice thickness of 1.25 mm, and a rotation time of 0.5 s per revolution. First, a plain scan was performed, followed by a contrast-enhanced scan. The contrast agent used was ioversol injection (350 mg L/ml), administered through the median cubital vein at a rate of 2.0-2.6 ml/s.

MRI examination

Fifty patients did not undergo MRI examination for special reasons, while the remaining 247 underwent abdominal MRI plain scan, DWI, and contrast-enhanced scan. MRI data was acquired using Siemens Prisma 3.0T and Philips Archieva 3.0T magnetic resonance scanners with abdominal coils. The scanning sequences included: (1) Axial in-phase and out-of-phase T1WI: TR 5.16ms, TE 1.35ms, slice thickness 4.0 mm, FOV 284.4 mm, matrix 195 × 320. (2) Axial T2WI: TR 3600ms, TE 76ms, slice thickness 6 mm, FOV 284.4 mm, matrix 182 × 320. (3) Axial contrast-enhanced T1WI: TR 5.19ms, TE 1.35ms, slice thickness 4.0 mm, FOV 284.4 mm, matrix 182 × 320. (4) DWI sequence: TR 7000ms, TE 58ms, slice thickness 6 mm, FOV 308.8 mm, matrix 104 × 128. First, a plain scan was performed, followed by a contrast-enhanced scan. The contrast agent used was gadoxetic acid disodium injection, administered through the median cubital vein at a dose of 0.1 ml/kg body weight over 12 s, followed by a flush of 20 ml of saline. Arterial phase, portal venous phase, and equilibrium phase images were acquired with delays of 20 s, 60 s, and 3 min, respectively.

Criteria for assessing pulmonary metastasis and definitions of other imaging signs

Pulmonary metastasis is deemed to have occurred if the following two criteria are met: (1) Pathological confirmation of pulmonary lesions following surgery. (2) A comprehensive assessment based on the following criteria: (a) Presence of a definite primary pulmonary lesion of HAE, as illustrated in Fig. 2; (b) Positive immunological test results for HAE, such as Enzyme-Linked Immunosorbent Assay (ELISA); (c) Pulmonary lesions exhibiting typical CT findings (e.g., multiple soft tissue density nodules of varying sizes in the peripheral zones of both lungs, often accompanied by cavitation and calcification), consistent with the imaging manifestations described by Eroglu et al.¹⁸, as depicted in Fig. 3.

Calcification signs: Punctate calcified particles can be seen within the lesion, accompanied by flocculent or irregularly large calcified foci. Cavity signs: The larvae proliferate into large lesions, with internal coagulation and vascular occlusion, which further lead to localized ischemic necrosis and liquefaction into irregular gelatinous substances, sometimes presenting crab claw-like liquefied cavities¹⁹. Marginal zone: The lesion has an unclear margin, showing infiltrative growth, rich in neovascularization and inflammatory cells; DWI shows a high signal area, and dynamic enhancement shows mild enhancement. Microvesicles: Small honeycomb or cystic structures are present within and at the edge of the lesion, showing high signal intensity on both T2WI and MRCP. Bile duct invasion: Hepatic bile ducts are infiltrated by lesions or have unclear boundaries with lesions, dilation, or bile duct obstruction²⁰. Vascular invasion: Defined as morphologically recognizable vascular lumen infiltration, no contrast agent filling after vascular occlusion, or associated thrombosis²¹.

Image analysis and clinical data

Two radiologists with ten years of diagnostic experience independently analyzed the abdominal and chest CT scans. In the event of any disagreement between the two radiologists regarding the interpretation of pulmonary metastasis on the chest CT, a chief physician with over 30 years of experience in imaging diagnosis made the final determination. The collected imaging features encompass the number, size, and type of lesions, PNM stages, extent of hepatic involvement, as well as the presence or absence of calcification signs, cavity signs, enhancement, marginal zone, microvesicles, bile duct dilation, bile duct invasion, hepatic gate invasion, hepatic vein invasion, inferior vena cava invasion, hepatic artery invasion, portal vein invasion, and other organ metastases. In multiple lesions, the lesion size refers to the maximum diameter of the most significant lesion. The intraclass correlation coefficient(ICC) was used to evaluate the consistency of measurements between the two doctors. An ICC greater than 0.75 indicated good consistency, and the average measurement value of the two doctors was then taken for further analysis. In addition, the clinical characteristics of the patients were collected, including sex, age, ethnic group, exposure history to cattle and sheep, HBV five markers, treatment methods, and complications.

Establishment and evaluation of prediction models

In this study, a nomogram model was constructed using the training dataset. Univariate and multivariate logistic regression analyses were employed to screen imaging and clinical indicators for predicting HAE pulmonary metastasis, and the resulting independent predictors were used to build the nomogram model. The model’s predictive ability was evaluated using the receiver operating characteristic (ROC) curve, and the area under the curve (AUC), as well as sensitivity and specificity, were calculated based on the maximum value of the Youden index. Simultaneously, calibration curves (CRC) were employed to evaluate the model’s goodness of fit, ensuring its stable and reliable performance in the test set. In addition, to more comprehensively evaluate the clinical value of the prediction model, we introduced decision curve analysis (DCA) to quantify the net benefit to patients across different threshold probabilities. This approach was used to assess the clinical utility of the prediction model and select the optimal model.

Statistical analysis

Statistical analysis was performed using RStudio 4.3.2, Medcalc 20.022, and SPSS 25. In SPSS 25, the Shapiro-Wilk test was first used to check the normality of continuously distributed data. Those conforming to the normal distribution were expressed as Mean ± SDs, and independent sample t-tests were used to compare the two groups. Categorical variables are expressed in frequencies, and comparisons between the two groups are made using the chi-square test. Univariate analysis was performed using independent sample t-tests or chi-square tests, and influencing factors with p < 0.05 were selected for inclusion in multivariate logistic regression analysis. ROC curves were drawn using Medcalc. The nomogram and calibration curves were plotted using the rms package in R, while decision curves were created with the rmda package. A p-value of less than 0.05 was considered statistically significant.

Source link