Specificity For Malignant Lesions Research Articles

Application of an artificial intelligence-based system in the diagnosis of breast ultrasound images obtained using a smartphone

BackgroundBreast ultrasound (US) is useful for dense breasts, and the introduction of artificial intelligence (AI)-assisted diagnoses of breast US images should be considered. However, the implementation of AI-based technologies in clinical practice is problematic because of the costs of introducing such approaches to hospital information systems (HISs) and the security risk of connecting HIS to the Internet to access AI services. To solve these problems, we developed a system that applies AI to the analysis of breast US images captured using a smartphone.MethodsTraining data were prepared using 115 images of benign lesions and 201 images of malignant lesions acquired at the Division of Breast Surgery, Gifu University Hospital. YOLOv3 (object detection models) was used to detect lesions on US images. A graphical user interface (GUI) was developed to predict an AI server. A smartphone application was also developed for capturing US images displayed on the HIS monitor with its camera and displaying the prediction results received from the AI server. The sensitivity and specificity of the prediction performed on the AI server and via the smartphone were calculated using 60 images spared from the training.ResultsThe established AI showed 100% sensitivity and 75% specificity for malignant lesions and took 0.2 s per prediction with the AI sever. Prediction using a smartphone required 2 s per prediction and showed 100% sensitivity and 97.5% specificity for malignant lesions.ConclusionsGood-quality predictions were obtained using the AI server. Moreover, the quality of the prediction via the smartphone was slightly better than that on the AI server, which can be safely and inexpensively introduced into HISs.

Utility of Fine-Needle Aspiration Cytology in Diagnosis of Lymphadenopathy

Introduction: Abnormal shape, size, and consistency of the lymph node constitute lymphadenopathy. The etiology of lymphadenopathy varies from reactive to malignancy. It is the significant manifestation of local and systemic ailments specially an occult malignancy. Fine-needle aspiration cytology (FNAC) is a simple, reliable initial diagnostic modality. Materials and Methods: This is a retrospective study from January 2005 to December 2017. The lesions were classified into reactive lymphadenopathy, granulomatous lymphadenitis, necrotizing lymphadenitis, suppurative lymphadenitis, lymphoma, and metastatic lesions. Primary organ was identified in available cases of metastatic lesions. Cytological and histopathological correlation was done in available cases. Results: A total of 3676 cases were retrieved constituting 16% of total FNAC performed during the study, females outnumbered the male. Most of the cases were observed in the age of 21–30 years. Cervical lymph nodes were the common site of involvement (88.8%). Among the benign lesions, granulomatous (40%) was most common followed by reactive (39%). Metastatic lymphadenopathy (13%) was the most common malignant lesion. Squamous cell carcinoma was the most common metastatic tumor followed by adenocarcinoma. The lung was the most common primary organ. FNAC is having high sensitivity and specificity for malignant lesions than the benign. Conclusion: FNAC is a rapid, initial diagnostic tool for lymphadenopathy. It is an effective tool to differentiate inflammatory, benign, and malignant lesions.

Multiparametric MRI Evaluation of Complex Ovarian Masses

ObjectiveTo assess the role of diffusion-weighted imaging and dynamic contrast-enhanced magnetic resonance imaging in the categorization of complex ovarian masses into benign and malignant. Materials and MethodsThis prospective study was done on 33 complex ovarian masses. T1 and T2-weighted sequences, diffusion-weighted imaging, apparent diffusion coefficient, and dynamic contrast-enhanced magnetic resonance imaging were performed on 1.5 T MRI. Time-intensity curves, tissue signal intensity on unenhanced T1 images (SI0), maximum absolute contrast enhancement (SImax), time to reach SImax (Tmax), maximum relative SI (SIrel = [SImax − SI0]/SI0 ×100), maximum Slope (Slopemax = SIrel/Tmax ×100), and wash in rate (WIR = [SImax − SI0]/Tmax) were calculated. Histopathological diagnosis was taken as gold standard. ResultsA total of 20/33 masses were benign, 2/33 were borderline tumors, and 11/33 were malignant. Diffusion restriction was seen in all malignant masses and 13/20 benign masses. The mean apparent diffusion coefficient values showed a significant difference between malignant and benign, with 81.8% sensitivity and 63.6% specificity. Type III curve showed 100% specificity for malignant lesions. Tmax and Slopemax were useful in differentiating benign and malignant masses; with Tmax cut-off at 73.5 seconds having a high specificity (81.8%) and Slopemax cut-off at 0.83%/s having high sensitivity (91%) and negative predictive value (94.4%). ConclusionMultiparametric MRI confers high diagnostic accuracy in stratifying complex ovarian masses.

Computed tomography findings of focal lung lesion and their pathological correlations

Background: Focal lung masses are common radiological fi ndings which need to be characterized to plan management. Suspicious lesions require invasive investigations and aggressive treatment. However, likely benign lesions require no invasive investigations.Objectives: We aimed to correlate the computerized tomography fi ndings of focal lung lesions with pathological diagnoses.Methods: A total of 56 patients were enrolled in the study. Plain and contrast enhanced computerized tomography scan of chest was done in spiral computerized tomography. Radiologist’s computerized tomography reports were obtained. Pathological diagnosis were followed in all cases and recorded. Computerized tomography reports were compared with pathological reports and validity of computerized tomography in predicting the nature of lung lesion was assessed.Results: Overall, squalors cell carcinoma was the most common histological cell type (46.34%) followed by adenocarcinoma (41.46%). In 95.12% of total cases, size of the malignant lesion was more than 3cm. Squalors cell Carcinoma was the most common tumor located centrally (71.42%) and adenocarcinoma peripherally (55.55%). Spiculated margin was most commonly associated in malignant lesions. Among morphological characters, chest wall invasion and meditational invasion were statistically significant for malignancy. Contrast enhancement more than 20 hounsfield unit had more than 80% sensitivity and specificity for malignant lesions. Sensitivity of computerized tomography combining all morphological characteristics for depicting malignancy was 97.6% and specificity was 46.7%.Conclusion: The present study showed that contrast enhanced computerized tomography provides more precise tumor localization, characterization and intrathoracic extension to differentiate benign from malignant lung lesion. Enhancement of the lesion is more specific for malignancy than other morphological characteristic.Journal of Kathmandu Medical College, Vol. 4, No. 2, Issue 12, Apr.-Jun., 2015, page: 59-64

Four-Year Results of Low-Dose CT Screening and Nodule Management in the ITALUNG Trial

Recruitment and nodule management are critical issues of lung cancer screening with low-dose computed tomography (LDCT). We report subjects' compliance and results of LDCT screening and management protocol in the active arm of the ITALUNG trial. Three thousand two hundred six smokers or former smokers invited by mail were randomized to receive four annual LDCT (n = 1613) or usual care (n = 1593). Management protocol included follow-up LDCT, 2-[18F]fluoro-2-deoxy-D glucose positron emission tomography (FDG-PET), and CT-guided fine-needle aspiration biopsy (FNAB). One thousand four hundred six subjects (87%) underwent baseline LDCT, and 1263 (79%) completed four screening rounds. LDCT was positive in 30.3% of the subjects at baseline and 15.8% subsequently. Twenty-one lung tumors in 20 subjects (1.5% detection) were found at baseline, and 20 lung tumors in 18 subjects (0.5% detection) in subsequent screening rounds. Ten of 18 prevalent (55%) and 13 of 17 incident (76%) non-small-cell cancers were in stage I. Interval growth enabled diagnosis of lung cancer in 16 subjects (42%), but at least one follow-up LDCT was obtained in 741 subjects (52.7%) over the screening period. FDG-PET obtained in 6.5% of subjects had 84% sensitivity and 90% specificity for malignant lesions. FNAB obtained in 2.4% of subjects showed 90% sensitivity and 88% specificity. Positivity of both FDG-PET and FNAB invariably predicted malignancy. Surgery for benign lesions was performed on four subjects (10% of procedures) but followed protocol violations on three subjects. High-risk subjects recruited by mail who entered LDCT screening showed a high and stable compliance. Efficacy of screening is, however, weakened by low detection rate and specificity. Adhesion to management protocol might lessen surgery for benign lesions.

Palpable Breast Lumps: Fine-Needle Aspiration Cytology versus Histopathology: a Correlation of Diagnostic Accuracy

Objective: To study the correlation between diagnostic accuracy of FNAC and histopathology in palpable breast lumps. Methods: A prospective study of 50 patients was conducted at K. J. Somaiya Medical College, Research Center and Hospital, Mumbai. All female patients were randomly selected, irrespective of their age, religion, marital status, occupation or social status. Every patient underwent a FNAC done on OPD basis by a trained pathologist from the Pathology Department in K. J. Somaiya College, Research Center and Hospital, following a thorough clinical check-up. Every patient subjected to FNAC underwent a definitive surgical procedure. All specimens so obtained were subjected to histopathology. The results thus obtained from histopathology were matched with those of FNAC and a correlation was sought based on statistical tests. Results: Results of all patients were collected and tabulated. Statistical analysis was performed on the tabulated data and sensitivity and specificity with positive and negative predictive value were obtained. Of all 50 patients selected, only 2 were false negative for malignancy i.e., they were detected as benign lesions on FNAC but found to be malignant on histopathology. Hence, the sensitivity of the study was 96% and the specificity for malignant lesions was 100%. The positive predictive value was 100% and the negative predictive value was 95.12% Conclusion: FNAC is a cheap, rapid, reliable and accurate test of diagnosing a palpable breast lump if done by an expert and trained pathologist. It gives a good histological correlation. It may obviate the need for another surgical biopsy prior to definitive surgery for malignancy.

CT-guided Transthoracic Needle Biopsy with an 18-Guage Automated Cutting Needle: Diagnostic benefits and safety

Purpose: We wanted to evaluate the diagnostic benefits and safety of performing CT-guided transthoracic needle biopsy (TTNB) with using an18-gauge automated cutting needle. Materials and Methods: We retrospectively reviewed the medical records of 332 patients who underwent CT-guided transthoracic biopsy with an automated 18-gauge cutting needle for 341 thoracic lesions. The histopathologic results, diagnostic accuracy and complication rate were determined. Results: 316 (92.7%) of 34l lesions were histopathologically diagnosed by CT-guided TTNB. The lesions were malignant in 172 patients and benign in 144 patients. The diagnostic accuracy was 92.9%. The sensitivity and specificity for malignant lesions were 93% and 92.3%, respectively. The sensitivity and specificity for benign lesions were 92.3% and 93%, respectively. TTNB induced complications developed for 54 of 341 (15.8%) procedures. The most frequent complication was pneumothorax (41/341, 12%). Nine patients with pneumothorax (9/41, 22%) required chest tube placement. Conclusion: CT-guided TTNB with an 18-gauge automated cutting needle could be useful for making the histopathologic diagnosis of malignant and benign pulmonary lesions as the technique provides a big enough tissue sample for histopathologic examination. The procedure-related complication rate was acceptably low.

Fine Needle Aspiration of Focal Lung Lesion under CT Fluoroscopic Guidance

Purpose: We wanted to assess the efficacy of using CT fluoroscopy during Fine Needle Aspiration (FNA) for focal lung lesion. Materials and Methods: We retrospectively reviewed 22 patients who had undergone FNA under CT fluoroscopic guidance. The final diagnosis of focal lung lesion was based on surgery, the results of biopsy taken from another site or the clinical diagnosis with imaging follow-up. We reviewed the imaging and recorded the location of the lesion, the lesion size and the depth from the puncture site. Results: In 24 cases of 22 patients, 16 lesions were benign and 8 were malignant. The sensitivity and specificity for malignant lesions were 100%, and 75%, respectively. The mean size of the malignant lesions was 3.751.68 cm, and that of the benign lesion was 3.052.98 cm. The mean depth of the lesion divided by the size was 1.85±1.09 for 20 lesions of the correct result group and 3.131.73 for 4 lesions of the incorrect result group; the difference between the two groups was statistically different (p=0.030). The complications after FNA were pneumothoraces in three cases and parenchymal hemorrhage in one case, but no significant complication was noted after the procedures. Conclusion: We could perform the FNA accurately and safely under CT fluoroscopy guidance, so this technique can be used for FNA of focal lung lesion for obtaining a correct result with fewer complications.

Reply to Guermazi et al: “Are time‐intensity curves helpful in differentiating between benign vertebral compression fracture and vertebral metastasis?”

We appreciate the opportunity to respond to the comments of Guermazi et al regarding our recent article (1) in the Journal of Magnetic Resonance Imaging. The main results of the studies of Kopka et al (2) and Earls et al (3) were that the signal-to-noise and contrast-to-noise ratios were higher by using a power injector. We do accept that the peak enhancement percentage and slope in a contrast-enhanced dynamic MR study could be slightly different by using different injection methods. However, in our study, the enhancement percentage and slope showed no significant difference between the compression fracture and metastatic groups. It is the enhancement pattern (type A∼E) and especially the second part of the curve after the rapidly rising first part that is helpful in differentiating between benign and malignant lesions. In our results, the type D pattern (the early washout phenomenon) had a 100% specificity for malignant lesions. There is no evidence that a power injector will change the tissue enhancement pattern in comparison with a manual bolus injection. There are several MR perfusion studies using the manual bolus injection method (4-7). In those studies, Gd-DTPA was injected by hand within 2–4 seconds. Tsukamoto et al evaluated the marrow perfusion of the femoral head in a dog model by using manual injection of the contrast medium, which showed good correlation with a microsphere blood flow study (r = 0.86) (4). In Verstraete's study (5), the correlation coefficient between enhancement slope and vascularization was 0.93; therefore, manual bolus injection could also be an acceptable method for evaluating tissue blood perfusion in those institutions that do not have a power injector. We apologize for not picking up the mismatch type D case number in the final proofreading. There should be 13 type D cases, all belonging to the malignant group (groups III and IV) (the 18th line at the right aspect of page 311 was mistyped). The small number of cases limits our statistical significance; however, in combining our experience with other studies (6, 7, 8), we believe that an early washout phenomenon is a specific perfusion pattern for malignant lesions, which is a useful sign in equivocal cases. Tiffany Ting-Fang Shih M.D.*, Wei-Tsung Chen M.D. , * Department of Medical Imaging, National Taiwan University Hospital, Taipei, Taiwan, Department of Radiology, Taipei Municipal Jen-Ai Hospital, Taipei, Taiwan.

Tumor-associated glycoprotein distribution detected by monoclonal antibody B72.3 in salivary neoplasia.

The expression of tumor-associated glycoprotein (TAG-72), an oncofetal mucin-like tumor-associated glycoprotein derived from membrane-enriched fractions of metastatic breast carcinoma, has been detected by monoclonal antibody (MoAb) B72.3 in adenocarcinomas of breast, colon, lung, endometrium, pancreas, and ovary. The authors reported the scope of TAG-72 expression detected by MoAb B72.3 in salivary neoplasia. They examined 96 salivary lesions (53 malignant and 37 benign primary tumors, 2 metastatic carcinomas, and 4 other benign lesions) and 17 normal tissues from parotid glands and found: diffuse TAG-72 expression in 29 of 55 (53%) malignant tumors and 6 of 36 (17%) benign tumors and in no normal tissue; focal TAG-72 expression in 10 of 55 (17%) malignant salivary tumors, 10 of 37 (25%) benign salivary tumors (all benign mixed tumors), and 1 of 17 (6%) histologically normal parotid gland ducts. Any expression of TAG-72, whether diffuse or focal, was found to have a 71% sensitivity for detecting salivary malignant tumors, but an unacceptably low specificity for malignant lesions (57%). Alternatively, if only diffuse TAG-72 expression was regarded as indicative of malignancy, the specificity of diffuse TAG-72 expression was 86%, but sensitivity of detection decreased to 53%. The authors studied a subset of benign and malignant mixed tumors (BMT and MMT) and found that 12 of 15 (80%) MMT diffusely and strongly expressed TAG-72, 2 of 15 MMT (13%) expressed TAG-72 focally, and 1 MMT (7%) was nonreactive. By contrast, most BMT did not express TAG-72; only sparse, focal TAG-72 expression was seen in 10 of 27 (37%) BMT. If diffuse TAG-72 expression is considered indicative of malignancy, its sensitivity and specificity for malignant mixed tumors is 80% and 100%, respectively. The authors suggest that diffuse TAG-72 expression may resolve conflicts in determining whether or not a mixed tumor is malignant.