Search

Objectives: To develop radiomics-based classifiers for preoperative prediction of fibrous capsule invasion in renal cell carcinoma (RCC) patients by CT images., Methods: In this study, clear cell RCC (ccRCC) patients who underwent both preoperative abdominal contrast-enhanced CT and nephrectomy surgery at our hospital were analysed. By transfer learning, we used base model obtained from Kidney Tumour Segmentation challenge dataset to semi-automatically segment kidney and tumours from corticomedullary phase (CMP) CT images. Dice similarity coefficient (DSC) was measured to evaluate the performance of segmentation models. Ten machine learning classifiers were compared in our study. Performance of the models was assessed by their accuracy, precision, recall, and area under the receiver operating characteristic curve (AUC). The reporting and methodological quality of our study was assessed by the CLEAR checklist and METRICS score., Results: This retrospective study enrolled 163 ccRCC patients. The semiautomatic segmentation model using CMP CT images obtained DSCs of 0.98 in the training cohort and 0.96 in the test cohort for kidney segmentation, and DSCs of 0.94 and 0.86 for tumour segmentation in the training and test set, respectively. For preoperative prediction of renal capsule invasion, the AdaBoost had the best performance in batch 1, with accuracy, precision, recall, and F1-score equal to 0.8571, 0.8333, 0.9091, and 0.8696, respectively; and the same classifier was also the most suitable for this classification in batch 2. The AUCs of AdaBoost for batch 1 and batch 2 were 0.83 (95% CI: 0.68-0.98) and 0.74 (95% CI: 0.51-0.97), respectively. Nine common significant features for classification were found from 2 independent batch datasets, including morphological and texture features., Conclusions: The CT-based radiomics classifiers performed well for the preoperative prediction of fibrous capsule invasion in ccRCC., Advances in Knowledge: Noninvasive prediction of renal fibrous capsule invasion in RCC is rather difficult by abdominal CT images before surgery. A machine learning classifier integrated with radiomics features shows a promising potential to assist surgical treatment options for RCC patients., (© The Author(s) 2024. Published by Oxford University Press on behalf of the British Institute of Radiology. All rights reserved. For permissions, please email: journals.permissions@oup.com.)

Objective: To develop a radiomic model based on low-dose CT (LDCT) to distinguish invasive adenocarcinomas (IAs) from adenocarcinoma in situ /minimally invasive adenocarcinomas (AIS/MIAs) manifesting as pure ground-glass nodules (pGGNs) and compare its performance with conventional quantitative and semantic features of LDCT, radiomic model of standard-dose CT, and intraoperative frozen section (FS)., Methods: A total of 147 consecutive pathologically confirmed pGGNs were divided into primary cohort (43 IAs and 60 AIS/MIAs) and validation cohort (19 IAs and 25 AIS/MIAs). Logistic regression models were built using conventional quantitative and semantic features, selected radiomic features of LDCT and standard-dose CT, and intraoperative FS diagnosis, respectively. The diagnostic performance was assessed by area under curve (AUC) of receiver operating characteristic curve, sensitivity, and specificity., Results: The AUCs of quantitative-semantic model, radiomic model of LDCT, radiomic model of standard-dose CT, and FS model were 0.879 (95% CI, 0.801-0.935), 0.929 (95% CI, 0.862-0.971), 0.941 (95% CI, 0.876-0.978), and 0.884 (95% CI, 0.805-0.938) in the primary cohort and 0.897 (95% CI, 0.768-0.968), 0.933 (95% CI, 0.815-0.986), 0.901 (95% CI, 0.773-0.970), and 0.828 (95% CI, 0.685-0.925) in the validation cohort. No significant difference of the AUCs was found among these models in both the primary and validation cohorts (all p > 0.05)., Conclusion: The LDCT-based quantitative-semantic score and radiomic signature, with good predictive performance, can be pre-operative and non-invasive biomarkers for assessing the invasive risk of pGGNs in lung cancer screening., Advances in Knowledge: The LDCT-based quantitative-semantic score and radiomic signature, with the equivalent performance to the radiomic model of standard-dose CT, can be pre-operative predictors for assessing the invasiveness of pGGNs in lung cancer screening and reducing excess examination and treatment.

Objective: To investigate the effect of reducing pixel size on the consistency of radiomic features and the diagnostic performance of the downstream radiomic signatures for the invasiveness for pulmonary ground-glass nodules (GGNs) on CTs., Methods: We retrospectively collected the clinical data of 182 patients with GGNs on high resolution CT (HRCT). The CT images of different pixel sizes (0.8mm, 0.4mm, 0.18 mm) were obtained by reconstructing the single HRCT scan using three combinations of field of view and matrix size. For each pixel size setting, radiomic features were extracted for all GGNs and radiomic signatures for the invasiveness of GGNs were built through two modeling pipelines for comparison., Results: The study finally extracted 788 radiomic features. 87% radiomic features demonstrated inter pixel size variation. By either modeling pipeline, the radiomic signature under small pixel size performed significantly better than those under middle or large pixel sizes in predicting the invasiveness of GGNs ( p 's value <0.05 by Delong test). With the independent modeling pipeline, the three pixel size bounded radiomic signatures shared almost no common features., Conclusions: Reducing pixel size could cause inconsistency in most radiomic features and improve the diagnostic performance of the downstream radiomic signatures. Particularly, super HRCTs with small pixel size resulted in more accurate radiomic signatures for the invasiveness of GGNs., Advances in Knowledge: The dependence of radiomic features on pixel size will affect the performance of the downstream radiomic signatures. The future radiomic studies should consider this effect of pixel size.

Objective: MRI is being used increasingly as a modality that can provide important information about breast cancer. Diffusion-weighted imaging (DWI) is an imaging technique from which apparent diffusion coefficient (ADC) values can be calculated in addition to obtaining important structural information which cannot be obtained from other imaging studies. We did not find any significant relationships between ADC values and prognostic factors, but did provide some explanations for conflicting results in the literature., Methods: The ADC results of 61 females with invasive ductal carcinomas were evaluated. DWI was performed and ADC values were calculated from the area in which restriction of diffusion was the highest in ADC mapping. B value was 500 and region of interest (ROI) was designated between 49 and 100 mm 2 . Calculations were performed automatically by the device. Tissue samples were obtained for prognostic factor evaluation. The relationships between ADC and prognostic factors were investigated. Comparisons between groups were made with one-way ANOVA and Kruskal Wallis test. Pairwise comparisons were made with Dunn's test. Analyses of categorical variables were made with Chi-square test., Results: We found a weak negative correlation between ADC and Ki-67 values (r = -0.279; p = 0.029). When we compared ADC values in regard to tumour type, we found no significant differences for tumour grade, Ki-67 positivity, estrogen receptor positivity, progesterone receptor positivity, C-erb B2, lymphovascular invasion and ductal carcinoma in situ or lobular carcinoma in situ component. On a side note, we found that mean ADC values decreased as tumour grade increased; however, this was not statistically significant., Conclusion: The literature contains studies that report conflicting results which may be caused by differences in B values, ROI area and magnetic field strength. Multicentre studies and systematic reviews of these findings may produce crucial data for the use of DWI in breast cancer. Advances in knowledge: To determine if any significant relationship exists between DWI findings and prognostic factors of breast cancer.

Objective: This study was undertaken to examine the characteristics of cancers detected at the earliest possible point on MRI and to determine their significance., Methods: This institutional review board-approved Health Insurance Portability and Accountability Act-compliant retrospective study evaluated invasive breast cancers ≤1 cm histologically. MRI was performed within 6 months before diagnosis. Between 1 January 2005 and 31 December 2015, 163 cancers in 161 patients were evaluated. Breast Imaging-Reporting and Data System lesion characteristics were assessed by two radiologists independently. In cases of disagreement, arbitration by a third reader was performed., Results: Cancers ≤1 cm became more obviously malignant as they enlarged with regard to shape (p = 0.021), margin (p = 0.0006), internal enhancement (p = 0.0158) and kinetics (p = 0.0001). Cancers ≤5 mm had benign characteristics of circumscribed margins in 71% (71/100), round/oval shape in 67% (67/100) and persistent enhancement in 41% (41/100). High T 2 signal was found in 17% (28/62), distributed equally among different sizes (p = 0.3920). In ≤5-mm cancers (59%, 12/29), a comparison study to show interval growth was more often needed to determine the need for biopsy. When interval growth determined biopsy, this was evident within 24 months and cancers remained node negative despite this delay., Conclusion: Benign characteristics are present in most invasive cancers ≤5 mm. Small cancers on MRI may need to demonstrate growth to determine need for biopsy. Advances in knowledge: MR lesion characteristics may not be helpful in determining whether small lesions on MR are benign or malignant. However, as 97% of cancers in our study showed interval change when a prior MR for comparison was available, new lesions or increasing size should lead to consideration of biopsy.

Objective: To investigate the application and value of fast low-angle shot three-dimensional (FLASH-3D) dynamic contrast-enhanced MRI for the pre-operative staging of endometrial carcinoma., Methods: This prospective study enrolled 48 patients with complete clinical data and pathologically confirmed endometrial carcinoma from July 2012 to March 2014. After routine MRI examination, subjects underwent FLASH-3D dynamic contrast-enhanced examination. The dynamically enhanced features of the uterine wall and tumours were analyzed. FLASH-3D pre-operative staging and findings in relation to myometrial invasion were compared with post-operative pathological results in a double-blind manner., Results: There were 48 cases of pathologically proven endometrial carcinoma, including 34 patients with Stage I (Stage Ia 22 cases and Stage Ib 12 cases), 9 with Stage II, 3 with Stage III and 2 with Stage IV. The staging accuracy for endometrial carcinoma was 81% (39/48) using FLASH-3D dynamic contrast-enhanced sequences. The sensitivity, specificity and accuracy for the determination of deep myometrial invasion were 84%, 90% and 88%, respectively. There was no significant difference compared with the results of post-operative pathology (p > 0.05)., Conclusion: FLASH-3D dynamic contrast-enhanced imaging may be valuable for the early diagnosis and pre-operative staging of endometrial carcinoma. Its high accuracy for assessing deep myometrial invasion makes FLASH-3D imaging an important tool for selecting the optimal therapeutic protocol and for prognosis estimation., Advances in Knowledge: FLASH-3D can significantly improve the accurate assessment of the depth of tumour invasion into the myometrium and may thus help to guide clinical surgical choices and post-operative evaluation. FLASH-3D is thus a promising technique for the routine examination of female pelvic tumours.

Objectives: Micro-CT is a promising modality to determine breast tumour size in three dimensions in intact lumpectomy specimens. We compared the accuracy of tumour size measurements using specimen micro-CT with measurements using multimodality pre-operative imaging., Methods: A tabletop micro-CT was used to image breast lumpectomy specimens. The largest tumour dimension on three-dimensional reconstructed micro-CT images of the specimen was compared with the measurements determined by pre-operative mammography, ultrasound and MRI. The largest dimension of pathologic invasive cancer size was used as the gold standard reference to assess the accuracy of imaging assessments., Results: 50 invasive breast cancer specimens in 50 patients had micro-CT imaging. 42 were invasive ductal carcinoma, 6 were invasive lobular carcinoma and 2 were other invasive cancer. Median patient age was 63 years (range 33-82 years). When compared with the largest pathologic tumour dimension, micro-CT measurements had the best correlation coefficient (r = 0.82, p < 0.001) followed by MRI (r = 0.78, p < 0.001), ultrasound (r = 0.61, p < 0.001) and mammography (r = 0.40, p < 0.01). When compared with pre-operative modalities, micro-CT had the best correlation coefficient (r = 0.86, p < 0.001) with MRI, followed by ultrasound (r = 0.60, p < 0.001) and mammography (r = 0.54, p < 0.001). Overall, mammography and ultrasound tended to underestimate the largest tumour dimension, while MRI and micro-CT overestimated the largest tumour dimension more frequently., Conclusion: Micro-CT is a potentially useful tool for accurate assessment of tumour dimensions within a lumpectomy specimen. Future studies need to be carried out to see if this technology could have a role in margin assessment., Advances in Knowledge: Micro-CT is a promising new technique which could potentially be used for rapid assessment of breast cancer dimensions in an intact lumpectomy specimen in order to guide surgical excision.

Renal angiomyolipoma is recognised as a benign hamartomatous lesion with no obvious malignant potential. However, the tumour may show extrarenal/perinephric extension at times. Rarely, the lesion may extend into the renal vein and inferior vena cava (IVC) indicating aggressive behaviour. We present a case of an angiomyolipoma of the kidney with sonographic, CT and MRI evidence of extension into the renal vein and IVC.

Whilst imaging of poor prognostic features in rectal cancers has assisted pre-operative treatment stratification, such features have yet to be evaluated in colonic cancers. This study aims to develop criteria for identifying poor prognostic features in colonic tumours and assess the accuracy of CT prediction against histopathology. Criteria were developed for predicting T-stage and N-stage, the presence of extramural vascular invasion and involvement of the retroperitoneal surgical margin (RSM). These criteria were tested on 33 patients with colonic cancer who underwent pre-operative high-resolution CT of their tumour. Two radiologists (Obs 1 and Obs 2) identified independently these poor prognostic features and the results were compared with the final histopathological results. Histological agreement and interobserver variation were calculated using the kappa test. Accuracy of CT prediction of tumour extension beyond muscularis propria was 82% (Obs 1) and 70% (Obs 2). Correct prediction of RSM involvement was 76% (95% confidence interval (CI): 57.8-88.9%) and 79% (95%CI: 61.1-91%) for Obs1 and Obs 2, respectively, with significant agreement between observers (kappa = 0.455, p = 0.050). Prognosis was correctly predicted using CT in 82% (95%CI: 61.5-81.2%) (Obs1) and 85% (95%CI: 68.1-94.9%) (Obs2) with moderate agreement (kappa = 0.459, kappa = 0.527, respectively) with histology. In conclusion, CT has potential as the imaging modality of choice in the pre-operative prediction of poor prognostic features in colonic cancers and could play a role in future treatment stratification.

The purpose of this study is to evaluate the diagnostic ability of ultrasound and define the sonographic features of symptomatic intraductal and invasive breast carcinoma. To achieve this the ultrasound features of 488 invasive carcinomas and 65 non-screening detected intraductal carcinomas were compared retrospectively. The features included size, AP/W (anteroposterior diameter/width) ratio, shape, margin, internal echogenicity, internal echotexture, posterior acoustic transmission, bilateral edge shadowing sign and calcifications. The sensitivity and specificity of the detection of calcifications by ultrasound in comparison with mammography were also studied. The accuracy of ultrasound diagnosis is 92.0% for invasive carcinoma of breast and 84.8% for intraductal carcinoma. Differentiation of ultrasound features of intraductal and invasive carcinoma can be based on the internal hypoechogenicity, loss of bilateral edge shadowing, posterior acoustic transmission, irregular shape and non-uniform internal echotexture with odds ratio of 0.3, 0.3, 0.4, 0.5 and 0.5, respectively. Internal echogenicity was the only significant differentiating factor on multiple logistic regression analysis. Non-comedo type ductal carcinoma in situ can be differentiated from comedo type by irregular shape with odds ratio of 0.3. The sensitivity, specificity and accuracy rate for the detection of calcifications in invasive carcinomas by ultrasound were 65.1%, 61.9% and 63.2%; in comedo type intraductal carcinoma 62.5%, 66.7% and 63.6%, and in non-comedo type intraductal carcinoma 30.0%, 86.7% and 64.0%, respectively. The ultrasound appearance of non-screening detected intraductal carcinoma is relatively isoechoic in comparison with invasive carcinoma. More than 60% of microcalcifications in comedo type intraductal carcinoma can be accurately demonstrated by ultrasound. However, the role of ultrasound in detecting symptomatic intraductal carcinoma warrants further study.

Ampullary carcinomas have a significantly higher resectability rate and better prognosis than other periampullary carcinomas, although the prognosis is poor with advanced disease. Accurate tumour staging is therefore important in surgical planning. Our objective was to evaluate the usefulness of, and problems associated with, endoscopic ultrasound (EUS) in the pre-operative staging of ampullary tumours. 35 patients with ampullary tumours were pre-operatively examined with EUS. The imaging results were compared with histopathological findings of the resected specimen according to the TNM staging classification. The overall accuracy of tumour (T) staging was 74% (26/35) for all tumours, and 67% (6/9), 71% (10/14) and 83% (10/12) respectively for T1, T2 and T3 tumours. The overall accuracy of nodal (N) staging was 63%. In diagnosing pancreatic invasion, EUS had an accuracy of 86% (30/35), a sensitivity of 83% (10/12), and a specificity of 87% (20/23). In conclusion, EUS provides an accurate method of evaluating the stage of ampullary tumours, especially infiltration into the pancreas. This modality is useful to surgeons in deciding on an appropriate therapeutic approach and in giving a prognosis.