Differential diagnosis of Crohn's disease and intestinal tuberculosis based on ATR-FTIR spectroscopy combined with machine learning.

BACKGROUND: Crohn's disease (CD) is often misdiagnosed as intestinal tuberculosis (ITB). However, the treatment and prognosis of these two diseases are dramatically different. Therefore, it is important to develop a method to identify CD and ITB with high accuracy, specificity, and speed. AIM: To develop a method to identify CD and ITB with high accuracy, specificity, and speed. METHODS: A total of 72 paraffin wax-embedded tissue sections were pathologically and clinically diagnosed as CD or ITB. Paraffin wax-embedded tissue sections were attached to a metal coating and measured using attenuated total reflectance fourier transform infrared spectroscopy at mid-infrared wavelengths combined with XGBoost for differential diagnosis. RESULTS: The results showed that the paraffin wax-embedded specimens of CD and ITB were significantly different in their spectral signals at 1074 cm-1 and 1234 cm-1 bands, and the differential diagnosis model based on spectral characteristics combined with machine learning showed accuracy, specificity, and sensitivity of 91.84%, 92.59%, and 90.90%, respectively, for the differential diagnosis of CD and ITB. CONCLUSION: Information on the mid-infrared region can reveal the different histological components of CD and ITB at the molecular level, and spectral analysis combined with machine learning to establish a diagnostic model is expected to become a new method for the differential diagnosis of CD and ITB.

Haematogenous Pneumonia Caused by Kocuria Kristinae in a Patient with a Central Venous Catheter.

Kocuria kristinae is a Gram-positive commensal bacterium, rarely responsible for infection in immunocompromised patients. A 29-year-old woman affected by intestinal pseudo-obstruction and requiring home parenteral nutrition, was hospitalised for fever and shivering during the infusion through a long-term central venous catheter (CVC). Blood cultures were positive for K. kristinae infection. At a chest CT scan, two partially cavitated nodular lesions were evidenced. Meropenem antibiotic therapy was used locally and systemically, resulting in catheter use restoration. A chest CT scan two months later at follow-up showed two centimetric, fibrotic and disventilatory areas replacing the previous nodular thickenings. Kokuria kristinae was responsible for haematogenous pulmonary involvement with excavated nodules, requiring a differential diagnosis. Moreover, in the case of a CVC infection, in addition to the risk of right endocarditis, haematogenous pneumonia must also be considered. LEARNING POINTS: Kocuria kristinae is a Gram-positive commensal bacterium, potentially responsible for infection.In the case of central venous catheter infection, in addition to the risk of right endocarditis, haematogenous pneumonia must also be considered.A differential diagnosis between bacterial and fungal infection is necessary (less, but to be suspected, neoplastic metastases).

Advanced Nanoencapsulation-Enabled Ultrasensitive Analysis: Unraveling Tumor Extracellular Vesicle Subpopulations for Differential Diagnosis of Hepatocellular Carcinoma via DNA Cascade Reactions.

Tumor-derived extracellular vesicles (tEVs) hold immense promise as potential biomarkers for the precise diagnosis of hepatocellular carcinoma (HCC). However, their clinical translation is hampered by their inherent characteristics, such as small size and high heterogeneity and complex environment, including non-EV particles and normal cell-derived EVs, which prolong separation procedures and compromise detection accuracy. In this study, we devised a DNA cascade reaction-triggered individual EV nanoencapsulation (DCR-IEVN) strategy to achieve the ultrasensitive and specific detection of tEV subpopulations via routine flow cytometry in a one-pot, one-step fashion. DCR-IEVN enables the direct and selective packaging of multiple tEV subpopulations in clinical serum samples into flower-like particles exceeding 600 nm. This approach bypasses the need for EV isolation, effectively reducing interference from non-EV particles and nontumor EVs. Compared with conventional analytical technologies, DCR-IEVN exhibits superior efficacy in diagnosing HCC owing to its high selectivity for tEVs. Integration of machine learning algorithms with DCR-IEVN resulted in differential diagnosis accuracy of 96.7% for the training cohort (n = 120) and 93.3% for the validation cohort (n = 30), effectively distinguishing HCC, cirrhosis, and healthy donors. This strategy offers a streamlined workflow and rapid assay completion and requires only small-volume serum samples and routine clinical devices, facilitating the clinical translation of tEV-based tumor diagnosis.

Radiomics model based on MRI to differentiate spinal multiple myeloma from metastases: A two-center study.

Purpose: Spinal multiple myeloma (MM) and metastases are two common cancer types with similar imaging characteristics, for which differential diagnosis is needed to ensure precision therapy. The aim of this study is to establish radiomics models for effective differentiation between them. Methods: Enrolled in this study were 263 patients from two medical institutions, including 127 with spinal MM and 136 with spinal metastases. Of them, 210 patients from institution I were used as the internal training cohort and 53 patients from Institution II were used as the external validation cohort. Contrast-enhanced T1-weighted imaging (CET1) and T2-weighted imaging (T2WI) sequences were collected and reviewed. Based on the 1037 radiomics features extracted from both CET1 and T2WI images, Logistic Regression (LR), AdaBoost (AB), Support Vector Machines (SVM), Random Forest (RF), and multiple kernel learning based SVM (MKL-SVM) were constructed. Hyper-parameters were tuned by five-fold cross-validation. The diagnostic efficiency among different radiomics models was compared by accuracy (ACC), sensitivity (SEN), specificity (SPE), area under the ROC curve (AUC), YI, positive predictive value (PPV), negative predictive value (NPY), and F1-score. Results: Based on single-sequence, the RF model outperformed all other models. All models based on T2WI images performed better than those based on CET1. The efficiency of all models was boosted by incorporating CET1 and T2WI sequences, and the MKL-SVM model achieved the best performance with ACC, AUC, and F1-score of 0.862, 0.870, and 0.874, respectively. Conclusions: The radiomics models constructed based on MRI achieved satisfactory diagnostic performance for differentiation of spinal MM and metastases, demonstrating broad application prospects for individualized diagnosis and treatment.

MNDA expression and its value in differential diagnosis of B-cell non-Hodgkin lymphomas: a comprehensive analysis of a large series of 1293 cases.

AIMS: MNDA (myeloid nuclear differentiation antigen) has been considered as a potential diagnostic marker for marginal zone lymphoma (MZL), but its utility in distinguishing MZL from other B-cell non-Hodgkin lymphomas (B-NHLs) and its clinicopathologic relevance in diffuse large B-cell lymphoma (DLBCL) are ambiguous. We comprehensively investigated MNDA expression in a large series of B-NHLs and evaluated its diagnostic value. METHODS: MNDA expression in a cohort of 1293 cases of B-NHLs and 338  cases of reactive lymphoid hyperplasia (RLH) was determined using immunohistochemistry and compared among different types of B-NHL. The clinicopathologic relevance of MNDA in DLBCL was investigated. RESULTS: MNDA was highly expressed in MZLs (437/663, 65.9%), compared with the confined staining in marginal zone B-cells in RLH; whereas neoplastic cells with plasmacytic differentiation lost MNDA expression. MNDA expression was significantly higher in mantle cell lymphoma (MCL, 79.6%, p = 0.006), whereas lower in chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL, 44.8%, p = 0.001) and lymphoplasmacytic lymphoma (LPL, 25%, p = 0.016), and dramatically lower in follicular lymphoma (FL, 5.2%, p < 0.001), compared with MZL. 29.6% (63/213) of DLBCLs were positive for MNDA. The cases in non-GCB group exhibited a higher rate of MNDA positivity (39.8%) compared to those in GCB group (16.3%) (p < 0.001), and MNDA staining was more frequently observed in DLBCLs with BCL2/MYC double-expression (50%) than those without BCL2/MYC double-expression (24.8%) (p = 0.001). Furthermore, there was a significant correlation between MNDA and CD5 expression in DLBCL (p = 0.036). CONCLUSIONS: MNDA was highly expressed in MZL with a potential utility in differential diagnosis between MZL and RLH as well as FL, whereas its value in distinguishing MZL from MCL, CLL/SLL is limited. In addition, MNDA expression in DLBCL was more frequently seen in the non-GCB group and the BCL2/MYC double-expression group, and demonstrated a correlation with CD5, which deserves further investigation. The clinical relevance of MNDA and its correlation with the prognosis of these lymphomas also warrant to be fully elucidated.

Imaging Findings of Calcifying Nested Stromal-Epithelial Tumour of the Liver: A Case Report and Literature Review.

BACKGROUND: A calcifying nested stromal-epithelial tumour (CNSET) is an erratic primary liver malignant tumour frequently discovered in young girls and females. Neither its pathogenesis nor its nosogenesis is clearly known. While principally indolent, infrequent tumours with aggressive clinical progression have been defined. This paper describes a CNSET case with rare clinical and imaging features. CASE PRESENTATION: A 17-year-old girl initially presented with enlarged lymph nodes near the main portal vein of the liver and a large liver tumour. Lesions were identified on the imaging findings obtained via positron emission tomography-computed tomography (CT) scanning, including an abnormal increase of heterogeneous glucose metabolism in the intrahepatic mass, with a maximum standardised uptake value of around 3.2. The CT imaging showed multiple dense shadows in the lesion, while the magnetic resonance imaging indicated a long T1 and a slightly longer T2. CONCLUSIONS: This study summarises the imaging features of CNSETs to provide a reference for diagnosing liver tumours. In addition, the literature on the topics covered was systematically reviewed.

INSM1 is a useful neuroendocrine marker to differentiate pancreatic serous cystadenoma from pancreatic well-differentiated neuroendocrine tumors in cytology and surgical specimens.

INTRODUCTION: Differentiating pancreatic serous cystadenoma (SCA) from well-differentiated neuroendocrine tumors (WDNETs) based on histomorphology is critical yet challenging, particularly in small biopsy samples. Our study aimed to examine the expression profile of INSM1 in cytologic and surgical resection specimens from pancreatic SCA to evaluate its potential as a discriminative marker against pancreatic WDNET. METHODS: We characterized INSM1 immunohistochemistry in 34 patients with pancreatic SCA, comprising 23 surgical resections and 11 cytology specimens. As a control, we used 28 cytology specimens from pancreatic WDNET. Clinical information was retrieved through a review of electronic medical records. RESULTS: All 11 pancreatic SCA cytology specimens and 15 of 23 pancreatic SCA surgical resections exhibited absent INSM1 immunostaining. Each of the remaining eight surgical resection specimens demonstrated 1 % immunoreactivity. In contrast, 27 out of 28 (96 %) pancreatic WDNET cytology specimens were positive for INSM1 immunostaining, with a median immunoreactivity of 90 % and a range of 30-90 %. Overall, INSM1 immunostains perform similarly to chromogranin and synaptophysin in pancreatic SCA. CONCLUSIONS: The results indicate that INSM1 immunohistochemistry staining may serve as a useful neuroendocrine marker to differentiate pancreatic SCA from pancreatic WDNET in clinical practice. To our knowledge, this represents the first large-scale study to evaluate INSM1 immunostaining in surgical and cytology specimens from pancreatic SCA.

Dynamic Contrast-Enhanced Ultrasonography Combined With LR-M Classification Criteria for Differentiating Malignant Liver Nodules at High Risk for Hepatocellular Carcinoma.

OBJECTIVE: We aimed to investigate the value of quantitative parameters derived from dynamic contrast-enhanced ultrasonography (DCE-US) and a combination of these quantitative parameters with the LR-M classification criteria in distinguishing hepatocellular carcinoma (HCC) nodules and non-HCC malignancies. METHODS: HCC and non-HCC malignant nodules were grouped using pathologic results, and each nodule was classified using CEUS LI-RADS 2017. Quantitative CEUS analysis of each nodule was performed using VueBox, and quantitative parameters were compared between the HCC and non-HCC groups. The diagnostic efficacy of the LR-5 category for HCC was analyzed using the LR-M classification criteria along with time-related quantitative parameters. RESULTS: Of the 190 malignant liver nodules, 137 and 53 were HCCs and non-HCC malignancies, respectively. The median values of quantitative parameters RT (rise time), TTP (time to peak), mTTl (mean transit time local), and FT (fall time) in the non-HCC malignant group were lower than those in the HCC group, with p < 0.05. There was a statistically significant difference in WiAUC (wash-in area under the curve), WoAUC (wash-out area under the curve), WiWoAUC (wash-in and wash-out area under the curve), and WoR (wash-out rate) values between HCC and non-HCC malignant groups, with p < 0.05. Using LR-M washout time <60 s and FT ≤21.2 s as the new diagnostic standard, the LR-5 category showed a sensitivity of 83.9%, specificity of 96.2%, and positive predictive value of 98.3% for HCC diagnosis. CONCLUSION: DCE-US can facilitate the distinction of HCCs and non-HCC malignancies. Non-HCC malignancies present with earlier peak enhancement and more rapid and marked washout than HCC nodules. The combination of the LR-M classification criteria and FT ≤21.2 s can significantly improve the diagnostic sensitivity of the LR-5 category for HCC.

High-grade salivary carcinomas: A current insight on diagnostic pathology and the key to clinical decision making.

High-grade carcinomas of the salivary glands are a group of several tumor entities with highly malignant histologic appearances, and have an aggressive biological behavior accompanied by poor a prognosis. In general, they require more intensive treatment than low- or intermediate-grade carcinomas. High-grade salivary carcinomas are rare and the microscopic features often overlap between different tumor types, making an appropriate diagnosis challenging in daily practice settings. However, with recent rapid advances in molecular pathology and molecular-targeted therapy in this field, there is a growing need to properly classify tumors, rather than just diagnosing the cases as "high-grade carcinomas". This leads to specific treatment strategies. In this article, we review representative high-grade salivary gland carcinomas, including salivary duct carcinoma and its histologic subtypes, high-grade mucoepidermoid carcinoma, solid-type adenoid cystic carcinoma, and high-grade transformation of low- or intermediate-grade carcinomas, and discuss their differential diagnoses and clinical implications. Other rare entities, such as neuroendocrine carcinoma, NUT carcinoma, and metastatic carcinoma, should also be considered before diagnosing high-grade carcinoma, NOS. Of these tumors, salivary duct carcinoma has received the most attention because of its strong association with androgen deprivation and anti-HER2 therapies. Other tumor-type-specific treatments include anti-TRK therapy for high-grade transformation of secretory carcinoma, but further therapeutic options are expected to be developed in the future. It should be emphasized that detailed histological evaluation with adequate sampling, in addition to the effective use of molecular ancillary tests, is of the utmost importance for a suitable diagnosis.

Application of machine learning in the preoperative radiomic diagnosis of ameloblastoma and odontogenic keratocyst based on cone-beam CT.

OBJECTIVES: Preoperative diagnosis of oral ameloblastoma (AME) and odontogenic keratocyst (OKC) has been a challenge in dentistry. This study uses radiomics approaches and machine learning (ML) algorithms to characterize cone beam computed tomography (CBCT) image features for the preoperative differential diagnosis of AME and OKC and compares ML algorithms to expert radiologists to validate performance. METHODS: We retrospectively collected the data of 326 patients with AME and OKC, where all diagnoses were confirmed by histopathologic tests. A total of 348 features were selected to train six ML models for differential diagnosis by a five-fold cross-validation. We then compared the performance of ML-based diagnoses to those of radiologists. RESULTS: Among the six ML models, XGBoost was effective in distinguishing AME and OKC in CBCT images, with its classification performance outperforming the other models. The mean precision, recall, accuracy, F1-score, and area under the curve (AUC) were 0.900, 0.807, 0.843, 0.841, and 0.872, respectively. Compared to the diagnostics by radiologists, ML-based radiomic diagnostics performed better. CONCLUSIONS: Radiomic-based ML algorithms allow CBCT images of AME and OKC to be distinguished accurately, facilitating the preoperative differential diagnosis of AME and OKC. ADVANCES IN KNOWLEDGE: ML and radiomic approaches with high-resolution CBCT images provide new insights into the differential diagnosis of AME and OKC.

Differential diagnosis of thyroid orbitopathy - diseases mimicking the presentation or activity of thyroid orbitopathy.

Thyroid orbitopathy (TO) is the most common cause of orbital tissue inflammation, accounting for about 60% of all orbital inflammations. The inflammatory activity and severity of TO should be diagnosed based on personal experience and according to standard diagnostic criteria. Magnetic resonance imaging (MRI) of the orbit is used not only to identify swelling and to differentiate inflammatory active from non-active TO, but also to exclude other pathologies, such as orbital tumours or vascular lesions. However, a group of diseases can mimic the clinical manifestations of TO, leading to serious diagnostic difficulties, especially when the patient has previously been diagnosed with a thyroid disorder. Diagnostic problems can be presented by cases of unilateral TO, unilateral or bilateral TO in patients with no previous or concomitant symptoms of thyroid disorders, lack of symptoms of eyelid retraction, divergent strabismus, diplopia as the only symptom of the disease, and history of increasing diplopia at the end of the day. The lack of visible efficacy of ongoing immunosuppressive treatment should also raise caution and lead to a differential diagnosis of TO. Differential diagnosis of TO and evaluation of its activity includes conditions leading to redness and/or swelling of the conjunctiva and/or eyelids, and other causes of ocular motility disorders and eye-setting disorders. In this paper, the authors review the most common diseases that can mimic TO or falsify the assessment of inflammatory activity of TO.

Contrast-enhanced CT in the differential diagnosis of bladder cancer and paraganglioma.

PURPOSE: We sought to summarize the value of contrast-enhanced computed tomography (CECT) in the differential diagnosis of bladder paraganglioma (BPG) and bladder cancer. METHODS: The medical records of 19 patients with BPG (13 males, 6 females) and 56 patients with bladder cancer (49 males, 7 females) between November 2007 and June 2023 were retrospectively reviewed. All patients underwent unenhanced and contrast-enhanced CT scanning. RESULTS: Patient age (46.4 ± 11.1 years vs. 58.6 ± 16.0 years), tumor calcification (1/19 vs. 18/56), stalk (0/19 vs. 10/56), internal vessels (15/19 vs. 19/56) and the enlarged adjacent supplying artery (14/19 vs. 10/56) were significantly different between BPG and bladder cancer (P < 0.05). The CT value in the corticomedullary phase (92.4 ± 16.6 HU vs. 64.0 ± 14.5 HU) and the contrast-enhanced value in the corticomedullary phase (54.5 ± 17.4 HU vs. 28.5 ± 12.8 HU) were significantly greater in BPG patients than in bladder cancer patients (P < 0.001), with corresponding area under the curve values of 0.930 and 0.912, respectively. The optimal cutoff values were 83.2 HU and 38.5 HU, respectively. A CT value > 83.2 HU in the corticomedullary phase and a contrast-enhanced CT value > 38.5 HU in the corticomedullary phase were used to indicate BPG with sensitivities of 78.9% and 89.5%, respectively, and specificities of 94.6% and 75.0%, respectively. CONCLUSION: The corticomedullary phase of CECT plays an important role in the preoperative differential diagnosis of BPG and bladder cancer.

Pseudomyogenic Hemangioendothelioma Arising in the Maxillary Sinus, an Uncommon and Misdiagnosed Tumor: A Case Report.

Pseudomyogenic hemangioendothelioma (PMHE) is a rare vascular neoplasm with an intermediate to low-grade malignant potential. Only 5% of PMHEs occur in the head and neck. This tumor exhibits different histological patterns and mimics other vascular tumors, myoid tumors, or carcinomas. The distinction between these tumors can be a very challenging situation for pathologists. In this article, we report the first case, to our knowledge, of PMHE arising in the maxillary sinus, to highlight this uncommon entity and discuss differential diagnoses.

Atypical Presentation of Bronchogenic Cyst in the Retroperitoneal Space.

Bronchogenic cysts, benign congenital malformations resulting from abnormal tracheobronchial tree budding, primarily manifest in the mediastinum, with retroperitoneal occurrence being exceedingly rare. Typically incidental findings on imaging, and their diagnosis pose challenges, particularly when malignancy is suspected. We present a case involving a 55-year-old woman diagnosed with chronic back pain. Physical examination revealed a painful mass in the left renal region. Subsequent MRI identified a smooth mass in the left adrenal gland without infiltration of surrounding structures. Laparoscopic surgery successfully removed the lesion without complications. Pathomorphological examination confirmed a gelatinous-filled cyst, identified as a retroperitoneal bronchogenic cyst in the left adrenal gland. Increasing reports of retroperitoneal bronchogenic cysts contribute to a better understanding of their characteristics, aiding preoperative diagnosis. However, given potential malignancy and definitive diagnosis through histopathological examination, surgical resection remains the preferred method.

Complete excision of a Skene gland cyst mimicking cystocele.

Skene gland abnormalities include skene gland cysts and abscess. These can be differentiated from urethral diverticulum based on clinical findings. The exact incidence of Skene gland abnormalities is unknown as they are relatively rare. They are usually seen in middle-aged female patients but have recently been reported in newborn girls. We present a video case of a large, adult-onset Skene gland cyst, which was evaluated based on clinical findings, radiological aspects and histopathological findings. The differential diagnosis was carried out step-by-step in order to avoid sequelae and complete excision was performed in order to achieve optimal results, both for long-term functional and anatomical outcomes.

A case report of multicentric reticulohistiocytosis with atypical cutaneous presentation.

Multicentric reticulohistiocytosis (MRH) is a rare systemic disorder characterized by histiocytic hyperplasia that mainly involves the skin, mucous membranes, and joints. The typical clinical features include papules, nodules, and arthritis. MRH lesions are relatively extensive but small and scattered. Joint inflammation is characterized by diffuse symmetric polyarthritis as the first symptom, which can be severe and disabling due to destructive joint changes. MRH is easily misdiagnosed in clinical practice. Here, we report the case of an elderly male patient who presented with polyarticular pain in the hip and interphalangeal joints as the first manifestation, followed by the development of large, isolated, bulging skin nodules, which are atypical MRH lesions. This is rare in all MRH case reports, and we made the correct diagnosis by combining skin histopathology, immunohistochemistry, and other clinical examinations. We performed surgical treatment on the local skin lesions of this patient. This case suggests that clinicians should actively correlate the condition and accurately diagnose MRH when encountering atypical skin changes or other diseases as the first symptom and explore the mechanisms of MRH and other clinical manifestations.

Diagnostic Performance of Endoscopic Ultrasound Elastography for Differential Diagnosis of Gallbladder Polyp.

BACKGROUND AND AIMS: It is difficult to differentiate between neoplastic and non-neoplastic gallbladder (GB) polyps before surgery. Endoscopic ultrasound-elastography (EUS-EG) is a non-invasive complementary diagnostic method. The utility of EUS-EG in the differential diagnosis of GB polyps has not been investigated. We aimed to investigate the diagnostic performance of EUS-EG for the differential diagnosis of GB polyps. METHODS: Patients with GB polyps were prospectively enrolled from June 2020 until November 2022. EUS-EG and semi-quantitative evaluation of the strain ratio (SR) were performed for differential diagnosis of GB polyps. Fifty-three eligible patients were divided into two groups based on the final diagnosis after surgery. Patient demographics, EUS characteristics, and SR values were compared. Receiver-operating characteristic (ROC) curve analysis was performed to determine the optimal cutoff SR value that discriminates between neoplastic and non-neoplastic GB polyps. RESULTS: The median SR value for neoplastic polyps (32.93 [interquartile range: 22.37-69.02]) was significantly higher than for non-neoplastic polyps (5.40 [2.36-14.44]; p<0.001). There were significant differences in SR values between non-neoplastic, benign neoplastic (23.38 [13.62-39.04]), and malignant polyps (49.25 [27.90-82.00]). The optimal cut-off SR value to differentiate between neoplastic and non-neoplastic polyps was 18.4. In multivariable logistic regression, SR value >18.4 (odds ratio 33.604, 95% confidence interval 2.588-436.292) was an independent predictor of neoplastic polyps. CONCLUSIONS: EUS-EG and SR values can be used as a supplementary method for evaluating GB polyps.

[Translated article] Comparative study of deep lipomas and atypical lipomatous tumours: Malignancy risk factors.

BACKGROUND AND OBJECTIVES: The diagnostic suspicion of an atypical lipomatous tumour (ALT) is difficult. The aim of this study is to delve into the most controversial diagnostic aspects of the subject. MATERIAL AND METHOD: Observational, longitudinal and retrospective study of a series of 96 deep adipose tumours (75 lipomas and 21 TLA) from 2006 to 2016: demographic, clinical, imaging and pathological variables were analysed and compared, as well as other variables related to treatment and oncological outcomes of the patients. A descriptive analysis of the collected variables was performed for the statistical study. To evaluate the potential predictor variables of malignancy, a multivariate logistic regression analysis was performed, including those that were statistically significant in the univariate analysis. RESULTS: Older age at diagnosis, lower limb location and larger size were significantly more frequent in ALTs. MRI findings showed no statistically significant differences between the two groups. In multivariate analysis, the same clinical variables were confirmed as predictors of malignancy. In the ROC curve, an optimal cut-off point of 134.0mm was used as a predictor of malignancy. CONCLUSIONS: Advanced age, location in the lower limbs and larger size are risk factors for malignancy in the differential diagnosis of deep lipomas and atypical lipomatous tumours. No radiological variable on MRI reached significance as a predictor of malignancy in our series.