CT Features Research Articles

Machine learning techniques for diagnosis of rare diseases from medical images

Introduction/Background: Rare diseases are life-threatening or chronically debilitating conditions that affect a small percentage of the population. Early and accurate diagnosis of rare diseases is challenging due to their complexity and limited understanding. Conventional diagnostic methods are often inadequate, time-consuming, and expensive. Machine learning (ML) has emerged as a promising technique for the analysis of medical images to support the diagnosis of various diseases including rare disorders. ML algorithms can learn complex patterns from large medical imaging datasets to aid clinicians in disease diagnosis, prognosis, and detection of complications. Materials and Methods: A structured search was conducted in electronic databases, including PubMed, Scopus, Web of Science, and Google Scholar to identify peer-reviewed articles published between 2013 to 2023 related to ML-based diagnosis of rare diseases from medical images. A total of 187 articles were identified after removing duplicates. The titles and abstracts of retrieved articles were screened to determine their relevance based on the research objective. Finally, 51 full-text articles were selected for the final review. The key data extracted from selected studies included diseases, imaging modalities, ML algorithms, performance metrics, datasets, and limitations. Results: Most studies evaluated deep learning-based ML techniques, with convolutional neural networks (CNN) being the most applied algorithm. CNNs were mainly used to classify rare diseases based on specific imaging features in X-rays, CT, and MRI scans. Diseases frequently investigated included cardiovascular, neurological, and rare genetic disorders. Publicly available datasets such as Deep Lesion, MSD, and ChestX-ray14 were commonly utilized. Most studies reported high classification accuracy ranging from 80-95% on test datasets. However, limited generalizability due to small private datasets and lack of external validation were limitations. Discussion: The results demonstrate the potential of deep learning for the image-based diagnosis of rare diseases. Features learned from large imaging data sets enabled CNNs to distinguish subtle abnormalities. Hybrid models combining CNNs with other techniques like recurrent neural networks further improved performance. Overall, ML showed promise as a decision support tool. However, more multi-center validation studies are needed using larger and diverse datasets before clinical adoption. Standardization of performance metrics and reporting is also required to establish reliability and generalizability. Conclusion: This comprehensive review provided insights into ongoing research applying ML to medical images for rare disease diagnosis. While initial results are encouraging, further work is still necessary before ML can be reliably used in clinical decision making. Larger collaborative efforts and open-source datasets are needed to advance the field. Standardization of ML frameworks can also promote reproducible research and comparison of different methodologies.

Synergizing traditional CT imaging with radiomics: a novel model for preoperative diagnosis of gastric neuroendocrine and mixed adenoneuroendocrine carcinoma

ObjectiveTo develop diagnostic models for differentiating gastric neuroendocrine carcinoma (g-NEC) and gastric mixed adeno-neuroendocrine carcinoma (g-MANEC) from gastric adenocarcinoma (g-ADC) based on traditional contrast enhanced CT imaging features and radiomics features.MethodsWe retrospectively analyzed 90 g-(MA)NEC (g-MANEC and g-NEC) patients matched 1:1 by T-stage with 90 g-ADC patients. Traditional CT features were analyzed using univariable and multivariable logistic regression. Tumor segmentation and radiomics features extraction were performed with Slicer and PyRadiomics. Feature selection was conducted through univariable analysis, correlation analysis, LASSO, and multivariable stepwise logistic. The combined model incorporated clinical and radiomics predictors. Diagnostic performance was assessed with ROC curves and DeLong’s test. The models’ diagnostic efficacy was further validated in subgroup of g-NEC vs. g-ADC and g-MANEC vs. g-ADC cases.ResultsTumor necrosis and lymph node metastasis were independent predictors for differentiating g-(MA)NEC from g-ADC (P < 0.05). The clinical model’s AUC was 0.700 (training) and 0.667(validation). Five radiomics features were retained, with the radiomics model showing AUC of 0.809 (training) and 0.802 (validation). The combined model’s AUCs were 0.853 (training) and 0.812 (validation), significantly outperforming the clinical model (P < 0.05). Subgroup analysis revealed that the combined model exhibited acceptable performance in differentiating g-NEC from g-ADC and g-MANEC from g-ADC, with AUC of 0.887 and 0.823 in the training cohort and 0.852 and 0.762 in the validation cohort.ConclusionA combined model based on traditional CT imaging and radiomic features provides a non-invasive and effective preoperative diagnostic method for differentiating g-(MA)NEC from g-ADC.

Characterization of exclusive rib lesions detected by [68Ga]Ga-PSMA-11 PET/CT.

The objective of this study was to characterize exclusive costal lesions detected by 68Gallium-labelled prostate-specific membrane antigen ([68Ga]Ga-PSMA-11) PET/computed tomography (CT) at initial staging or biochemical recurrence (BCR) in prostate cancer (PCa) patients, and to identify clinical and/or PET/CT criteria associated with benign and malignant lesions. We retrospectively identified 54 patients with PCa who underwent [68Ga]Ga-PSMA-11 PET/CT for initial staging (N = 39) or BCR (N = 15) and whose reports described rib lesions, at the exclusion of any other lesions, whether doubtful, suspicious, or established. Posttherapy prostate-specific antigen (PSA) levels were used to determine whether those lesions were benign or malignant. Each patient's prostate-specific membrane antigen PET/CT report was classified as true positive, true negative, false positive, or false negative based on the posttherapy PSA level. We then assessed whether any clinical and/or PET/CT criteria could help differentiate benign from malignant lesions, and if any criteria were misleading. Among the 54 patients, 46 (85.2%) had 64 benign costal lesions, and eight (14.8%) had 10 malignant lesions. PET/CT reports indicated rib lesions as benign/equivocal in 38/54 (55.6%) patients and malignant in 16/54 (29.6%). Benign features on CT were the only parameter significantly associated with the final diagnosis. Factors such as patient age, maximum standardized uptake value of lesions, lesion dispersion, and malignant features described on CT were found to be misleading when deciding the malignant or benign status. Most exclusive costal lesions detected by [68Ga]Ga-PSMA-11 PET/CT are benign. Apart from specific benign CT features, no clinical or PET/CT criteria reliably differentiate benign from malignant costal lesions.

CT Features of Fibrous Dysplasia

CT Features of Fibrous Dysplasia

Dual Sulfone-Bridged Triphenylamine Heteroaromatics for High-Performance Blue Organic Electroluminescence.

Polycyclic heteroaromatics (PHAs) are a highly versatile class of functional materials, especially applicable as efficient luminophores in organic light-emitting diodes (OLEDs). Those constructed by tethered phenyl surrounding the main group center attract extensive attention due to their excellent OLED device performance. However, the development of such a class of emitters is often limited to boron, nitrogen-doped π-conjugated heterocycles. Herein, we proposed a novel kind of blue emitter by constructing a donor-acceptor molecular configuration, utilizing a dual sulfone-bridged triphenylamine (BTPO) core and mono/di-diphenylamine (DPA) substituents. The twisted D-A molecular structures and appropriate donor strength facilitate the effective separation of natural transition orbitals, endowing the emitters with charge-transfer dominant hybridized local and charge-transfer characteristics for the excited states. Both BTPO-DPA and BTPO-2DPA own small S1-T1 splitting energy, thus demonstrating blue thermally activated delayed fluorescence. The more symmetrical structure and enhanced CT features brought by additional DPA moiety confer BTPO-2DPA with a shorter delayed fluorescence lifetime, a higher fluorescence quantum yield and narrower emission. Therefore, BTPO-2DPA based OLED devices exhibit superior blue electroluminescence performance, with external quantum efficiencies reaching 12.31 %.

CT and MRI of pancreatic cystic lesions: tricks of the trade

PurposeTo describe CT and MRI findings of pancreatic cystic lesions.BackgroundThe growing use of ultrasound, CT and MRI has led to increasing incidental detection of pancreatic cystic lesions. These lesions range from benign entities (e.g., acute necrotic collection) to potentially malignant neoplasms (e.g., IPMN) and malignant neoplasms (e.g., ductal adenocarcinoma). CT and MRI can help to differentiate pancreatic cystic lesions.FindingsPancreatic cystic lesions neoplasms often have pathognomonic CT and MRI characteristics. It is important to consider size, number of cysts, septa, the relationship with the main pancreatic duct, internal nodules and wall enhancement. Guidelines from different societies have been developed to promote and standardize the management and follow-up of pancreatic cystic lesions.ConclusionCT and MRI play a crucial role in the detection and characterization of pancreatic cystic lesions.

Comparison of computed tomography findings between perianal tumors originating from the anal sacs and perianal glands in dogs.

Tumors originating from the anal sacs and perianal glands are common tumors of the perianal region. However, differentiation of perianal tumors based on CT findings remains challenging because perianal tumors can show overlapping CT characteristics. Therefore, this retrospective, multi-institutional study aimed to identify the CT findings that can differentiate between tumors originating from the anal sacs and perianal glands in canines. Twenty dogs were included, with nine apocrine gland anal sac adenocarcinomas (AGASACAs), six perianal gland adenomas, three perianal gland carcinomas, and two perianal gland epitheliomas. The tumor CT features included the tumor growth pattern, tumor shape, cystic changes within the tumor, contrast enhancement pattern, displacement of the anus, and identification of the anal sacs. The prevalences of suspected metastatic lymphadenopathy of the iliosacral lymph centers and lung metastases were evaluated. The tumor volumes and values normalized to body weight were also assessed. The tumor growth patterns differed significantly between AGASACAs and perianal gland tumors. AGASACAs mostly showed unilateral growth, whereas perianal gland tumors mostly exhibited bilateral growth. CT features, except for the tumor growth pattern, failed to show significant differences between tumors arising from the anal sacs and those arising from the perianal glands. Although histopathological diagnosis remains the gold standard, this study's findings suggest that CT analysis of tumor growth patterns may help differentiate AGASACAs from perianal gland tumors.

CT and MRI radiographic characteristics of seminal vesicle cysts

PurposeSeminal vesicle cysts are relatively uncommon benign lesions. Early and accurate diagnosis is crucial for guiding appropriate management strategies. The purpose of this study was to investigate the radiographic characteristics of seminal vesicle cysts in order to enhance the diagnostic proficiency of radiologists. MethodsWe conduct a comprehensive descriptive analysis of the CT and MRI findings pertaining to 15 patients with seminal vesicle cyst, complemented by a brief literature review. ResultThe mean age of 15 patients with seminal vesicle cyst was 57.53 ± 17.0 years. Among these patients, 53.33% (8/15) were incidentally detected, while 40% (6/15) presented with symptoms of abdominal or pelvic pain. One patient exhibited hematospermia. In 86.67% (13/15) of cases, the cysts were unilateral, while the remaining 13.33% (2/15) were bilateral. Radiological assessment comprised CT scans for 7 patients, MRI scans for 12 patients, and a combination of both modalities for 4 patients. Imaging characteristics revealed 11 cases of unilocular cysts and 4 cases of multilocular cysts, with the largest measuring approximately 16.0 × 9.8 cm. CT imaging distinctly demonstrated punctate or nodular calcification along the cyst wall in 3 cases. MRI, on the other hand, facilitated the visualization of cystic contents, including mucinous protein in 9 cases and hemorrhagic components in 2 cases. ConclusionThe radiographic manifestations of seminal vesicle cysts exhibit distinct characteristics, CT and MRI imaging modalities are instrumental in facilitating their accurate diagnosis.

Computed tomography of acute orbital trauma in children: a retrospective study

INTRODUCTION: Orbital injuries account for 36 to 64% of all blunt trauma to the facial bones. Diagnosis of orbital wall fractures is carried out using conventional radiography in several projections and computed tomography; magnetic resonance imaging is less commonly used. Often, due to the small displacement of orbital bone fragments, X-ray diagnosis is difficult. Multislice CT with reconstruction of sagittal, coronal projections and 3D in the bone window in such cases is the best visualization method. However, there are still no clear indications for the use of each method and there is no complete view of the CT characteristics of orbital damage.OBJECTIVE: To study the diagnostic value of computed tomography of orbital fractures in pediatric patients with acute trauma.MATERIAL AND METHODS: Retro- and prospectively analyzed the results of CT scans of 94 patients with facial fractures from 01/01/2023 to 09/01/2023 to assess combined orbital injuries. Orbital fractures were detected in 63 children (67.0%). There were 44 boys (69.84%), 19 girls (30.15%) aged from 7 months to 17 years 10 months (average age 10.3). Statistics: For statistical analysis, the method of calculating the nominal correlation using the Kramer coefficient (V) using the IBM SPSS Statistics software package was used.RESULTS: Concomitant injury was observed in 30 (47.6%); isolated 33 (52.36%). The number and frequency of fractures observed were as follows: superior wall of the orbit — 39 (61.9%); zygomatic complex — 19 (30.1%); lower wall of the orbit — 43 (68.2%); nasal bone — 17 (26.9%); lower jaw — 6 (9.5%); medial wall of the orbit — 28 (44.4%); upper jaw — 27 (42.8%); alveolar process — 6 (9.5%); isolated zygomatic arch — 14 (22.2%); Le Fort type I — 1 (1.5%); Le Fort type II — 2 (3.1%); and Le Fort type III — 0 (0%). Orbital hematomas were found in 25 patients (39.6%). It was revealed a strong correlation between the presence of contiguous skull fractures and multiple skull fractures (V=0.878, p<0.001), frontal sinus fractures and frontal sinus hemosinus (V=0.69, p<0.001).CONCLUSION: Orbital fractures are a frequent type of facial fracture occurring in children with blunt isolated and combined trauma. In children with blunt trauma, head CT combined with clinical examination is currently the optimal tool for diagnosing clinically significant orbital injuries. There is a strong correlation between orbital hematomas and orbital vault fractures, orbital hematomas and lattice fractures, adjacent skull and orbital vault fractures, and adjacent skull and lattice fractures. CT examination should include multiplanar (in coronal and sagittal planes) and 3D reconstructions, which help to assess the extent of the fracture, the presence of diastasis, the degree of muscle impingement. Based on the data obtained, the clinician determines the need for surgical intervention.

Prediction of benign and malignant ground glass pulmonary nodules based on multi-feature fusion of attention mechanism.

The purpose of this study was to develop and validate a new feature fusion algorithm to improve the classification performance of benign and malignant ground-glass nodules (GGNs) based on deep learning. We retrospectively collected 385 cases of GGNs confirmed by surgical pathology from three hospitals. We utilized 239 GGNs from Hospital 1 as the training and internal validation set, and 115 and 31 GGNs from Hospital 2 and Hospital 3, respectively, as external test sets 1 and 2. Among these GGNs, 172 were benign and 203 were malignant. First, we evaluated clinical and morphological features of GGNs at baseline chest CT and simultaneously extracted whole-lung radiomics features. Then, deep convolutional neural networks (CNNs) and backpropagation neural networks (BPNNs) were applied to extract deep features from whole-lung CT images, clinical, morphological features, and whole-lung radiomics features separately. Finally, we integrated these four types of deep features using an attention mechanism. Multiple metrics were employed to evaluate the predictive performance of the model. The deep learning model integrating clinical, morphological, radiomics and whole lung CT image features with attention mechanism (CMRI-AM) achieved the best performance, with area under the curve (AUC) values of 0.941 (95% CI: 0.898-0.972), 0.861 (95% CI: 0.823-0.882), and 0.906 (95% CI: 0.878-0.932) on the internal validation set, external test set 1, and external test set 2, respectively. The AUC differences between the CMRI-AM model and other feature combination models were statistically significant in all three groups (all p<0.05). Our experimental results demonstrated that (1) applying attention mechanism to fuse whole-lung CT images, radiomics features, clinical, and morphological features is feasible, (2) clinical, morphological, and radiomics features provide supplementary information for the classification of benign and malignant GGNs based on CT images, and (3) utilizing baseline whole-lung CT features to predict the benign and malignant of GGNs is an effective method. Therefore, optimizing the fusion of baseline whole-lung CT features can effectively improve the classification performance of GGNs.

Acute Splenic Sequestration: A Rare and Severe Complication of Sickle Cell Anemia in Adults: About a Case Report

Acute splenic sequestration is a well-known complication of sickle cell disorders, marked by a sudden drop in hemoglobin levels and significant, painless splenomegaly. This condition is typically observed in children with homozygous sickle cell disease. However, it is rarely reported in adults with heterozygous sickle cell conditions. We describe the case of a 28-year-old patient with a history of hemoglobin SC disease who suffered an acute splenic sequestration crisis. We review the CT characteristics of splenic sequestration, including splenic enlargement and an irregular peripheral rim of hypoenhancing or hypoechoic tissue, and discuss differential diagnoses. While acute splenic sequestration is predominantly a severe complication in children, timely diagnosis and treatment particularly red blood cell transfusions can lead to full recovery.

5179 Uncommon Benign Mimics Of Suspicious Adrenal Nodules

Abstract Disclosure: L. Zhu: None. L. Loh: None. V. Gan: None. P. Kek: None. Introduction: Adrenal incidentaloma is often detected in imaging studies. In addition to biochemical testing, imaging plays a key role in guiding the management of such lesions. We describe two uncommon cases of benign lesions masquerading as suspicious adrenal nodules. Case 1. A 54-year-old man with hypertension, diabetes, dyslipidaemia, ischaemic heart disease and atrial fibrillation (on warfarin) underwent CT scan to exclude malignancy when he presented with thigh DVT, and was found to have a right 3.2cm adrenal nodule. Adrenal hormone profile was unremarkable. CT contrast characterization showed attenuation of 30, 35, and 32 HU on pre-contrast, venous and 10-minute delayed phase respectively. Given the minimal enhancement in high density areas, possible haemorrhage was considered and he was put on imaging surveillance. CT scans over the following 5 years showed stability of the lesion, resolution of high-density areas, and pre-contrast attenuation of -35HU, suggestive of myelolipoma. Further follow-up over the next 2 years then revealed tumour enlargement to 8.5cm, progressive predominance of dense areas (58 HU), and development of compressive symptoms. In view of mass effect, growing heterogeneity and possibility of a collision tumour, right adrenalectomy was performed. Histopathology revealed myelolipoma with extensive haemorrhage and no evidence of malignancy. Case 2. A 65-year-old woman had hypertension, ischaemic heart disease, dialysis-dependent end-stage renal disease secondary to glomerulonephritis, and oncological history of left renal cell carcinoma (RCC) and low-grade papillary urothelial carcinoma, both surgically resected 10 years ago. A new 1.6cm right adrenal nodule was seen on her surveillance CT scan for RCC. It was not present in a CT scan 7 years ago. The nodule displayed suspicious features such as nodular avid rim enhancement, persistent enhancement on delayed phase of contrast washout and central cystic change. Biochemical screening revealed plasma free metanephrine of 0.33 nmol/L (NR: &amp;lt;0.33) and plasma free normetanephrine of 0.99 nmol/L (NR: &amp;lt;0.71). Other adrenal hormone profiles were unremarkable. Cystic phaeochromocytoma was suspected and she underwent right adrenalectomy and nephrectomy following alpha blockade. Histopathology revealed a well-circumscribed nodule in the supra-renal adipose tissue, with vascular proliferation and features of anastomosing haemangioma. The right kidney was reported to have acquired cystic disease associated RCC. She was well post-operatively and remained on follow-up with the urologist for her RCC. Discussion: Haemorrhagic adrenal myelolipoma and supra-renal anatomosing haemangioma are benign lesions that may have CT characteristics similar to those of sinister lesions. Pre-operative diagnosis can be challenging. Surgical excision may be required to confirm diagnosis and exclude malignancy. Presentation: 6/3/2024

Computed tomographic features of pulmonary and extrapulmonary lesions can be useful in prioritizing the diagnosis of hemangiosarcoma metastases in dogs.

The aim of this study was to evaluate the CT features of pulmonary metastases in dogs with hemangiosarcoma (HSA) at various sites. Additionally, the CT characteristics of extrapulmonary metastases in the same population were assessed. Retrospective, observational, and descriptive study conducted from April 2013 to January 2024. Dogs with histologically confirmed HSA and suspected or cytologically/histologically confirmed lung metastases were included. Dogs were excluded if they had a second primary tumor or only 1 unsampled pulmonary nodule. 33 dogs were included, with 26/33 [78.8%] having more than 10 metastatic pulmonary nodules. Most nodules were generalized (24/33 [72.7%]), miliary (29/33 [87.9%]) to subcentimetric (32/33 [97%]) in size, well-defined margins (29/33 [87.9%]), or a perilesional halo sign (24/33 [72.7%]). When more than 10 nodules were present, a generalized distribution was prevalent, while a peripheral location was more common when 2 to 10 nodules were present (P < .0001). In 32/33 (97%) cases, a pulmonary vessel was directly connected to the nodule (feeding vessel). After contrast administration, most lung metastases appeared homogenous (26/33 [78.8%]), although some showed areas of intense enhancement (5/33 [15.1%]) a feature also observed in extrapulmonary metastases with varying frequency (0% to 85.7%). Pulmonary HSA metastases were characterized by generalized, small (miliary/subcentimetric), well-defined nodules, commonly associated with a halo sign and feeding vessel. Intralesional areas of spotty postcontrast linear or amorphous strong hyperdensity were frequently observed especially in extrapulmonary metastases. These features may help radiologists and clinicians orient their diagnosis toward metastatic HSA.

Computed tomographic features of Morel-Lavallée lesions in 3 cats.

To describe the CT features of Morel-Lavallée lesions (MLLs) and their association with trauma in cats. This study aimed to investigate whether MLL fluid in cats accumulates between the dermis and deep fascia near superficial trauma, exhibits a vertical attenuation gradient, and shows no or only mild contrast enhancement on CT images. 3 young cats: a 1-year-old neutered male domestic shorthair cat, a 2-year-old neutered male domestic shorthair cat, and a 3-year-old neutered female British Shorthair cat. The clinical presentation of the 3 cats in the study involved recurrent swelling and fluid accumulation affecting at least 1 pelvic limb following trauma. In all cases, there was a fluid-filled space deep to the skin and superficial subcutaneous layer but superficial to the deep fascial plane overlying the muscular component of the pelvic limbs. The MLL fluid had a vertical attenuation gradient and no or only mild contrast enhancement. Morel-Lavallée lesions are often initially overlooked, likely due to distraction by concomitant significant bone injuries, and can be mistaken for hematomas, seromas, subcutaneous edema, contusions, or cellulitis, leading to less aggressive treatment. If left untreated, MLLs present a significant risk of extensive skin necrosis compared to other types of subcutaneous swelling.

Variability of radiological and clinical features in cases with usual interstitial pneumonia without honeycombing

BackgroundUsual interstitial pneumonia (UIP) cases without honeycombing (possible UIP) included various CT features and was often difficult to diagnose. PurposeThis study aimed to classify the cases with possible UIP on CT features using cluster analysis and evaluate the features of subsets of participants and the correlation of prognosis. Materials and MethodsThe study included 85 patients with possible UIP in the 2011 idiopathic pulmonary fibrosis (IPF) guideline with radiological diagnosis. All cases underwent surgical biopsies and were diagnosed by multidisciplinary discussion (MDD) from the nationwide registry in Japan. The readers evaluated pulmonary opacity, nodules, cysts, and predominant distribution which were reclassified by IPF guidelines in 2018. Additionally, cases were classified into four groups by cluster analysis based on CT findings. The differences in survival among IPF classification and the clusters were evaluated. ResultsCases were diagnosed as IPF (n = 55), NSIP (n = 4), unclassifiable (n = 23), and others (n = 3) by MDD. Cluster analysis revealed 4 clusters by CT features (n = 47, 16, 19 and 3, respectively). Cluster 1 had fewer lesions overall. Cluster 2 have many pure ground-glass opacities and ground-glass opacities with reticulation. Cluster 3 had many reticular opacities and nodules with few lower predominant distributions. Cluster 4 was characterized by peribronchovascular consolidation.The mean survival time of cluster 1 (4518 days) was significantly better than cluster 2, 3, and 4 (1843, 2196, and 1814 days, respectively) (p = 0.03). ConclusionIn conclusion, UIP without honeycombing included various CT patterns and MDD diagnoses. Significangly differences in prognosis were observed among clusters classified by CT findings.

Ranking the Relative Importance of Image Quality Features in CT by Consensus Survey

ObjectiveThis study sought to determine consensus opinions from subspecialty radiologists and imaging physicists on the relative importance of image quality features in CT. MethodsA prospective survey of subspecialty radiologists and medical physicists was conducted to collect consensus opinions on the relative importance of ten image quality features: axial sharpness, blooming, contrast, longitudinal sharpness, low contrast axial sharpness, metal artifact, motion, noise magnitude, noise texture, and streaking. The survey was first sent to subspecialty radiologists in volunteer leadership roles in the American College of Radiology and Radiological Society of North America, thereafter relying on snowball sampling. Surveyed subspecialties were abdominal, cardiac, emergency, musculoskeletal, neuro, pediatric, and thoracic radiology, and medical physics. Individual respondents’ ratings were normalized for calculation of mean normalized ratings and priority rankings for each feature within subspecialties. Also calculated were intraclass correlation coefficients across image quality features within subspecialties, and analysis of variance across subspecialties within each feature. ResultsMost subspecialties had moderate to excellent intraclass agreement. For every radiology subspecialty except musculoskeletal, motion was the most important image quality feature. There was agreement across subspecialties that axial sharpness and contrast are only moderately important. There was disagreement across subspecialties on the relative importance of noise magnitude. Blooming was highly important to cardiac radiologists, and noise texture was highly important to musculoskeletal radiologists. ConclusionImage quality preferences differ based on clinical tasks and challenges in each anatomical radiology subspecialty. CT image analysis and development of quantitative measures of quality and protocol optimization—and related policy initiatives—should be specific to radiology subspecialty.

Multicenter evaluation of CT deep radiomics model in predicting Leibovich score risk groups for non-metastatic clear cell renal cell carcinoma

BackgroundNon-metastatic clear cell renal cell carcinoma (nccRCC) poses a significant risk of postoperative recurrence and metastasis, underscoring the importance of accurate preoperative risk assessment. While the Leibovich score is effective, it relies on postoperative histopathological data. This study aims to evaluate the efficacy of CT radiomics and deep learning models in predicting Leibovich score risk groups in nccRCC, and to explore the interrelationship between CT and pathological features. Patients and MethodsThis research analyzed 600 nccRCC patients from four datasets, dividing them into low (Leibovich scores of 0–2) and intermediate to high risk (Leibovichscores exceeding 3) groups. Radiological models were developed from CT subjective features, and radiomics and deep learning models were constructed from CT images. Additionally, a deep radiomics model and a fusion model integrating radiological, radiomics, and deep learning features were introduced. Model performance was assessed by AUC values, while survival differences across predicted groups were analyzed using survival curves and the DeLong test. Moreover, the research investigated the connection between CT and pathological features derived from whole-slide pathological images in the second validation dataset. ResultsWithin the training dataset, four radiological, three radiomics, and thirteen deep learning features were selected to develop models predicting nccRCC Leibovich score risk groups. The deep radiomics model demonstrated superior predictive accuracy, evidenced by AUC values of 0.881, 0.829, and 0.819 in external validation datasets. Notably, significant differences in overall survival were observed among patients classified by this model (DeLong test p < 0.05 across all datasets). Furthermore, a correlation and complementarity were observed between deep radiomics features and pathological deep learning features. ConclusionsThe CT deep radiomics model precisely predicts nccRCC Leibovich score risk groups preoperatively and highlights the synergistic effect between CT and pathological data.

Differentiating primary from metastatic ovarian tumors of gastrointestinal origin by CT

PurposeTo determine differentiating CT imaging features of primary ovarian cancers from ovarian metastases of gastrointestinal origin. MethodsRetrospective study of 50 patients with new ovarian lesions on CT, half were primary ovarian cancers and half gastrointestinal metastases. Two blinded independent readers described tumor characteristics on CT (size, laterality, margin, etc.) and ancillary features (ascites, peritoneal seeding, lymphadenopathy, etc.). Patient age, sex, cancer history, and tumor marker levels for CA-125 and CEA were collected. Wilcoxon test and Pearson's chi-squared test were used for statistical analysis. Results50 patients with mean age of 62.1 years were included. Ovarian metastases were more likely to be cystic/mainly cystic (p=0.013), have smooth margins (p=0.011), and have no/mild enhancement (p<0.001). Primary ovarian lesions were associated with moderate to large volume of ascites (p=0.047) and more commonly seen with lymphadenopathy (p=0.008). Laterality was not significantly different between the two groups. CA-125 level was more commonly elevated in primary ovarian lesions (87% vs 50%, p=0.018), and with much higher values (1076.5 vs 155.1, p=0.013). CEA level was more commonly elevated in metastatic ovarian lesions (83.3% vs 15.4%, p<0.001), and with higher values (72.4 vs 2.1, p<0.001). ConclusionOvarian metastases were more frequently smooth-margined and cystic with little enhancement. Primary ovarian lesions were more commonly associated with lymphadenopathy and larger volume of ascites. Tumor markers CEA and CA-125 were more frequently elevated in metastatic and primary lesions, respectively. Cancer history was the only variable that increased the odds of metastasis and therefore it is important to always correlate with history of cancer.

MA08.09 Predictive CT Features of Visceral Pleural Invasion in Lung Cancer with Ground Glass Opacity

MA08.09 Predictive CT Features of Visceral Pleural Invasion in Lung Cancer with Ground Glass Opacity

Using CT features of cystic airspace to predict lung adenocarcinoma invasiveness.

Lung adenocarcinoma associated with cystic airspace (LACA) was once considered an uncommon manifestation of lung adenocarcinoma (LUAD), and understandings of it are limited; however, it is being observed more frequently in clinical practice. This study sought to assess the prevalence of LACA, and compare the high-resolution computed tomography (HRCT) features of LACA in patients with varying degrees of invasiveness. This study retrospectively reviewed the HRCT scans of 1,525 patients with LUAD ≤3 cm in diameter at the Shanghai Chest Hospital between January 2016 and May 2016. Each nodule was examined to detect the presence of cystic airspace. Additionally, we analyzed the qualitative HRCT findings of the cystic airspaces, including the pattern, number, wall component density, distribution, inner surface, mural nodules, septa, and vessels passing through the cystic airspace using the Pearson χ2 test or Fisher's exact test as appropriate. We also analyzed the quantitative measurements, such as the cystic airspace diameter, wall thickness, and thin-wall proportion, using a one-way analysis of variance or the Kruskal-Wallis rank-sum test as appropriate. LACAs were observed on HRCT in 11.5% (176/1,525) of the patients, of whom 7.1% (36/505) had pure ground-glass nodules, 13.5% (112/830) had mixed ground-glass nodules, and 14.7% (28/190) had solid nodules (P=0.001). The surgical procedures for LACAs varied (P=0.012). The incidence of LACAs increased as nodule diameter and invasiveness increased (both P<0.001). Statistically significant differences were observed in the wall component density, distribution, septa, vessels passing through the cystic airspace, cystic airspace diameter, wall thickness, and thin-wall proportion among the preinvasive lesion (PL), minimally invasive adenocarcinoma (MIA), and invasive adenocarcinoma (IAC) groups (P<0.001, P=0.024, P=0.001, P=0.025, P=0.001, P<0.001, and P<0.001, respectively). Wall component density increased as invasiveness increased (P<0.001). Unlike those in the MIAs and IACs, cystic airspaces in PLs typically lacked septa (P=0.001, and P<0.001, respectively). The IACs had larger cystic airspace diameters than the PLs (6.5 vs. 3.7 mm) (P<0.001). The IACs also had thicker wall thickness (11.8 vs. 6.8 mm, 11.8 vs. 8.3 mm) (P<0.001, and P<0.001, respectively) and smaller thin-wall proportions (181.5° vs. 264.8°, 181.5° vs. 223.8°) (P<0.001, and P=0.039, respectively) than the PLs and MIAs. The prevalence and characteristics of cystic airspaces on HRCT can be used to predict invasiveness in patients with LUADs ≤3 cm in diameter.