Accuracy For Diagnosis Of Malignancy Research Articles

A deep learning system to predict origins for cancers of unknown primary using cytological images from hydrothorax and ascites.

e13628 Background: Cancer of unknown primary (CUP) origin is still a mystery in the field of tumor diagnosis, which is difficult to be diagnosed by conventional imaging methods. The aim of this study is to develop a deep learning based method to assist pathologists in determining the origin of malignant hydrothorax and ascites using cytological images. Methods: By curating cytology images of hydrothorax and ascites from 42682 cases at three hospitals, we developed a deep-learning-based method -Tumor Origin diffeRentiation for Cytological Histology (TORCH)- for diagnosis of malignancy and prediction of tumor origin. We examined its performance on three individual test sets (n=12799). We compared TORCH with a panel of four pathologists on 496 cases that were randomly selected from test sets. Area under the receiver operating characteristic curve (AUROC) was used as the primary classification metric. Metrics including accuracy, sensitivity, specificity, precision and negative predictive value were used to measure the efficacy of TORCH versus pathologists. A score system was used to evaluate and compare the performance between our AI model and pathologist experts. Attention heatmap was used to interpret model prediction results. Results: TORCH model reached an overall micro-average one versus rest auroc of 0.957(95% CI: 0.955-0.959). On the three test sets, TORCH achieved an AUROC of 0.964 (95% CI 0.957-0.972), 0.947 (0.940-0.953), and 0.988 (0.985-0.993) in the diagnosis of cancer and 0.947 (0.942-0.952), 0.957 (0.953-0.960) and 0.972 (0.968-0.976) in localization of tumor origin, respectively. For primary tumor origin prediction, TORCH achieved an overall top-1 accuracy of 79.2% and top-3 accuracy of 98.5%. On three test sets, the top-1 accuracy was 75.7%, 79.4%, and 85.0% respectively. Compared with pathologists, TORCH demonstrated a better prediction efficacy (1.665 ± 0.677 vs. 1.265 ± 0.555; P <0.001). The mean diagnostic score of junior pathologists assisted with TORCH was also remarkably higher than the previous one (1.326 ± 0.728 vs. 1.101 ± 0.824; P <0.001). The comprehensive accuracy of attention heatmaps capturing main area of isolated tumor cells was 87.3% (95% CI: 81.9-92.8%). Conclusions: TORCH model was able to provide relatively accurate diagnosis of malignancy and prediction of tumor cell origin in hydrothorax and ascites. It can be used as an effective ancillary method to aid clinicians in allocating personalized therapy and improving patients’ prognosis.

Small Renal Masses without Gross Fat: What Is the Role of Contrast-Enhanced MDCT?

Increased detection of small renal masses (SRMs) has encouraged research for non-invasive diagnostic tools capable of adequately differentiating malignant vs. benign SRMs and the type of the tumour. Multi-detector computed tomography (MDCT) has been suggested as an alternative to intervention, therefore, it is important to determine both the capabilities and limitations of MDCT for SRM evaluation. In our study, two abdominal radiologists retrospectively blindly assessed MDCT scan images of 98 patients with incidentally detected lipid-poor SRMs that did not present as definitely aggressive lesions on CT. Radiological conclusions were compared to histopathological findings of materials obtained during surgery that were assumed as the gold standard. The probability (odds ratio (OR)) in regression analyses, sensitivity (SE), and specificity (SP) of predetermined SRM characteristics were calculated. Correct differentiation between malignant vs. benign SRMs was detected in 70.4% of cases, with more accurate identification of malignant (73%) in comparison to benign (65.7%) lesions. The radiological conclusions of SRM type matched histopathological findings in 56.1%. Central scarring (OR 10.6, p = 0.001), diameter of lesion (OR 2.4, p = 0.003), and homogeneous accumulation of contrast medium (OR 3.4, p = 0.03) significantly influenced the accuracy of malignant diagnosis. SE and SP of these parameters varied from 20.6% to 91.3% and 22.9% to 74.3%, respectively. In conclusion, MDCT is able to correctly differentiate malignant versus uncharacteristic benign SRMs in more than 2/3 of cases. However, frequency of the correct histopathological SRM type MDCT identification remains low.

Comparison of 22-gauge standard and Franseen needles in EUS-guided tissue acquisition for diagnosing solid pancreatic lesions: a multicenter randomized controlled trial

This large multicenter randomized controlled trial compared the diagnostic yields of 22-gauge standard and 22-gauge Franseen needles for EUS-guided tissue acquisition (EUS-TA) of solid pancreatic lesions. Consecutive patients with solid pancreatic lesions were prospectively randomized to EUS-TA using standard or Franseen needles. Samples obtained with the first needle pass and with second and subsequent passes were evaluated separately. The primary endpoint was the rate of accuracy for diagnosis of malignancy. Other endpoints were technical success rate, sample cellularity, adverse events, diagnostic accuracy in patient subgroups, and the diagnostic accuracy and numbers of second and subsequent needle passes. Of 523 patients undergoing EUS-TA, 260 were randomized to using standard 22-gauge needles and 263 to 22-gauge Franseen needles. The technical success rate in each group was 99.6%, with similar adverse event rates in the standard (1.5%) and Franseen (.8%) needle groups. First-pass EUS-TA using the Franseen needle resulted in significantly greater diagnostic accuracy (84.0% vs 71.2%, P< .001) and sensitivity (82.4% vs 66.7%, P< .001) than first-pass EUS-TA using a standard needle and also resulted in superior diagnostic accuracy in patients requiring immunostaining. Second and subsequent EUS-TA using Franseen needles showed significantly greater accuracy (94.7% vs 90.0%, P= .049) and sensitivity (94.0% vs 88.6%, P= .047) and required fewer needle passes (1.81 vs 2.03, P= .008) than using standard needles. EUS-TA with the Franseen needle is superior to EUS-TA with a standard needle with respect to diagnostic accuracy per pass, particularly in patients who require immunostaining, and number of passes when using macroscopic on-site evaluation. (Clinical trial registration numbers: UMIN000030634 and jRCTs052180062.).

Malignant Serous Effusions among Hospital In-patients in a Tertiary Care Hospital: A Descriptive Cross-sectional Study.

Cell block technique is an adjunct to conventional smears in the diagnosis of malignancy in effusion fluid. It aims at retrieving cellular material and concentrating cells in a small field, with preservation of cytomorphologic details. The objective of this study was to find the proportion of malignant serous effusions using cell block technique among hospital in-patients in a tertiary care centre. This was a descriptive cross-sectional study conducted among patients visiting a tertiary care centre between 1st June 2020 to 30th November 2020. Ethical approval was taken from the Institutional Review Committee (Reference number: 305202001). Using a convenience sampling method, 96 hospital in-patients were included in the study. Serous effusions were evaluated by conventional smears and cell block sections. Data was analysed using the Statistical Package for the Social Sciences version 23. Point estimate at 95% Confidence Interval was calculated along with frequency, percentage, mean and standard deviation. Among 96 hospital in-patients, 15 (15.62%) (8.35-22.88 at 95% Confidence Interval) were diagnosed as positive for malignancy by using cell block technique. By conventional smears, 80 (83.33%) cases turned out to be negative for malignancy, 13 (13.54%) were positive for malignancy and three (3.12%) were suspicious for malignancy. Of the three (3.12%) cases suspected for malignancy, two turned out to be positive for malignancy and one was found to be negative for malignancy on cell block technology. The proportion of malignant serous effusions was similar in comparison to other studies. Cell block technique could be routinely incorporated along with conventional smears for a more accurate diagnosis of malignancy on serous effusion.

Diagnostic efficacy, performance and safety of side-cut core needle biopsy for thyroid nodules: comparison of automated and semi-automated biopsy needles

This study was performed to compare the utility of the semi-automated and automated side-cut core biopsy needles for thyroid nodules. Between January 2014 and March 2020, biopsy was performed for 278 thyroid nodules using the semi-automated core needle and for 225 nodules using the automated core needle. Nondiagnostic rate, inconclusive rate, diagnostic performance and complication rates were evaluated and compared between two core needle types. There were 1.2% (6/503) nondiagnostic results and 15.9% (80/503) inconclusive results. Nondiagnostic rate between two needle types was not significantly different. The semi-automated type (33/278, 11.9%) showed lower inconclusive rate than the automated type (47/225, 20.9%) (p = 0.006). The sensitivity, specificity, PPV, NPV and diagnostic accuracy for diagnosis of malignancy of the semi-automated type were 70.18, 100, 100, 84.96 and 88.89%, respectively; the corresponding rates of automated type were 70.45, 100, 100, 86.6, and 89.84%. There were 12 minor complications: four hematomas (4/278, 1.4%) for the semi-automated type and eight hematomas (8/225, 3.6%) for the automated type, which difference was not statistically significant. Core needle biopsy for thyroid nodules using either the semi-automated or automated needle is a safe diagnostic tool. Semi-automated needle has lower inconclusive rate than automated needle.

Endoscopic ultrasound-guided fine-needle biopsy in patients with unexplained diffuse gastrointestinal wall thickening.

Background and study aims Endoscopic ultrasound-guided fine needle biopsy (EUS-FNB) is recommended after non-diagnostic biopsy in gastrointestinal wall thickening, although the performance of currently available FNB needles in this setting is unknown. We aimed to assess the diagnostic accuracy and safety of EUS-FNB and to evaluate the "T" wall staging in malignant pathology. Patients and methods This was a single center retrospective study that included all consecutive patients undergoing EUS-FNB for diffuse gastrointestinal wall thickening with at least one previous negative conventional endoscopic biopsy between January 2016 and November 2019. EUS-FNB was performed using linear-array echoendoscopes with slow-pull/fanning technique. Tissue acquisition was done with 19- or 22-gauge biopsy needles. Samples were included in formalin without rapid on-site evaluation and submitted for histopathological processing. The final diagnosis was based on conclusive histology or absence of evidence of disease progression after follow-up at least 6 months. Results Twenty-nine patients (21 men), with a median age of 68 (IQR: 56–77), were included. EUS-FNB was technically feasible and the sample quality was adequate for full histological assessment in all patients (100 %). Sensitivity, specificity, positive and negative predictive values, and overall accuracy for diagnosis of malignancy were 95.5 %, 100 %, 100 %, 83.3 %, and 96.3 %, respectively. In patients with malignant disease, the samples obtained allowed detection of signs of deep layer infiltration (“histological staging”) in 17 of 21 cases (81 %). No adverse events were noted. Conclusions The EUS-FNB technique demonstrated excellent diagnostic performance and safety in the study of unexplained diffuse gastrointestinal wall thickening. Histological staging was obtained in a high percentage of samples.

Intraoperative Frozen Section Biopsy of Uterine Smooth Muscle Tumors: A Clinicopathologic Analysis of 112 Cases With Emphasis on Potential Diagnostic Pitfalls.

Frozen sections of uterine smooth muscle tumors are infrequently required, and related diagnostic difficulties are seldom discussed. We analyzed the clinicopathologic features of 112 frozen sections of uterine smooth muscle tumors and determined the accuracy, reasons for deferrals, and causes of interpretational errors. Most patients (median age, 45 y) presented with pelvic mass symptoms (53%). The main reasons for a frozen section examination were an abnormal gross appearance including loss of the usual whorled pattern of leiomyoma (36 cases, 32.1%), and intraoperative discovery of an abnormal growth pattern and extrauterine extension of a uterine tumor (28 cases, 25%). There were 9 leiomyosarcomas and 103 leiomyomas, including 18 benign histologic variants. An accurate diagnosis of malignancy was achieved in all leiomyosarcomas, with the exception of a myxoid leiomyosarcoma. In 99 cases (88%), the frozen section diagnosis concurred with the permanent section diagnosis (false positives, 0.9%; false negatives, 0%). Misinterpretation of stromal hyalinization as tumor cell necrosis in a leiomyoma with amianthoid-like fibers was a major discrepancy. Two minor discrepancies did not lead to a change in management. The diagnosis was deferred in 10 cases (8.9%) because of stromal alterations, unusual cellular morphology, uncertain type of necrosis, and abnormal growth patterns. Thus, although various stromal and cellular alterations can cause diagnostic uncertainty, leading to deferrals, frozen section diagnosis of uterine smooth muscle tumors has a high accuracy rate. While a definitive frozen section diagnosis of malignancy may be made when there is unequivocal atypia, indisputable mitotic figures, and tumor cell necrosis, it is important to remember that nonmyogenic mesenchymal tumors may also mimic uterine smooth muscle tumors. In a frozen section setting, it would be sufficient to issue a diagnosis of "malignant mesenchymal tumor." For tumors that do not meet the criteria for malignancy, issuing a frozen section diagnosis of "atypical mesenchymal tumor and defer the histologic subtyping to the permanent sections" is appropriate.

Profile of anterior gradient 3 (AGR3) mRNA expression and serum levels in benign and malignant breast tumors.

Benign and malignant breast tumors are the most commonly diagnosed tumor in females. Early and accurate diagnosis of malignancy is essential for effective breast cancer treatment. Human anterior gradient 3 (AGR3) has been suggested as a potential biomarker for the early detection and prognostic determination of breast cancer. This study profiles AGR3 mRNA expression and serum protein levels in patients with benign and malignant breast tumors. A case-control study was conducted on 40 benign and 40 malignant breast tumor patients in Makassar, Indonesia. AGR3 mRNA and protein were detected using qRT-PCR and ELISA, respectively. This study found significantly higher AGR3 mRNA expression in benign than malignant breast tumors using qRT-PCR (p <0.001). In contrast, ELISA revealed no significant difference between AGR3 serum protein levels in benign and malignant breast tumors (p =0.507). AGR3 is associated with non-aggressive tumors and could be used as a marker for less aggressive breast tumors.

Can the height of the parotid tumor be a predictor of malignancy?

Accurate diagnosis of malignancy in the parotid gland before surgery is often challenging. Various clues should be used to increase the index of suspicion for malignancy. We hypothesized that malignant lesions of the parotid gland are located at the superior part of the gland compared to benign ones. A total of 169 consecutive patients were included in this study whose medical records were retrospectively reviewed. Benign and malignant tumors were compared in size, height difference from five anatomical landmarks: hard palate, mastoid tip, earlobe, condylar head, and mandibular notch. The cutoff heights from significant landmarks (hard palate, condylar head) were estimated with ROC analysis and chi-square test. Twenty-nine patients (17.2%) were diagnosed with malignant and 140 patients (82.8%) as benign. The height differed significantly between benign and malignant tumors when the reference point was set for the hard palate (P=0.024) and the condylar head (P=0.049), with the cutoff height from reference points to be 22.5, 51.5 mm, respectively. Diagnostic values of parotid level difference presented higher sensitivity (75.9% for hard palate, 72.4% for condylar head vs. 47.8% for fine needle aspiration cytology) and lower false negative rate (24.1% for hard palate, 27.6% for condylar head vs. 52.2% for fine needle aspiration cytology) compared to fine needle aspiration cytology. Malignant tumors of the parotid gland tend to locate at the superior part of the gland compared to benign tumors. Parotid tumors lying cephalad should raise an index of suspicion for malignancy. Height of the tumor in the parotid gland should be deliberately considered during the first encounter of the patient, which in turn could curate the next step in the diagnostic approach and treatment planning.

Contribution of Cell Block Obtained by Thoracentesis in the Diagnosis of Malignant Pleural Effusion.

Aim:The aim of this study wass to compare the cytological features of pleural exudative fluids by conventional smear (CS) method and cell block (CB) method and also to assess the utility of the combined approach for cytodiagnosis of these effusions.Materials and Methods:In all, 113 pleural exudative fluid samples were subjected to evaluation by both CS and CB methods over a period of 2 years. Cellularity, architecture patterns, morphological features, and yield for malignancy were compared, using the two methods. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy for diagnosis of malignancy were calculated for both the methods, using histology as a gold standard.Results:CB method provided higher cellularity, better architectural patterns, and additional yield for malignancy when compared with CS method. For 22 (40%) patients, histologic subtype was determined with CB especially for adenocarcinoma. The sensitivity, specificity, positive, and negative predictive values of cytology and CB were 48%, 100%, 100%, 67.8% and 59.2%, 100%, 100%, 72.8%, respectively.Conclusion:CB technique definitively increased detection of malignancy in pleural fluid effusion when used as an adjunct to CSs. Also, CB provides material suitable for molecular genetic analysis for targeted therapies especially in the treatment of adenocarcinoma.

Diagnostic Performance of Radial Probe Endobronchial Ultrasound without a Guide-Sheath and the Feasibility of Molecular Analysis.

BackgroundRadial probe endobronchial ultrasound (R-EBUS), is effective for tissue diagnosis of lung lesions. We evaluated the diagnostic performance of R-EBUS both a guide-sheath and fluoroscopy and identified factors associated with accurate diagnosis. The feasibility of molecular and genetic testing, using specimens obtained by R-EBUS, was also investigated.MethodsThe study retrospectively reviewed 211 patients undergoing R-EBUS without a guide-sheath and fluoroscopy, June 2016-May 2017. After excluding 27 patients of which the target lesion was not reached, 184 were finally included. Multivariate logistic regression was used, to identify factors associated with accurate diagnosis.ResultsAmong 184 patients, R-EBUS-guided biopsy diagnosed malignancy in 109 patients (59%). The remaining 75 patients (41%) with non-malignant results underwent additional work-ups, and 34 were diagnosed with malignancy. Based on final diagnosis, diagnostic accuracy was 80% (136/170), and sensitivity and specificity for malignancy were 76% (109/143) and 100% (27/27), respectively. In multivariate analysis, peripheral location (adjusted odds ratio [aOR], 3.925; 95% confidence interval [CI], 1.203–12.811; p=0.023), and central position of the probe (aOR, 2.435; 95% CI, 1.424–7.013; p=0.035), were associated with accurate diagnosis of malignancy. Molecular and genetic analyses were successful, in all but one case, with inadequate specimens.ConclusionR-EBUS-guided biopsy without equipment, is effective for tissue diagnosis. Peripheral location and central position of the radial probe, were crucial for accurate diagnosis. Performance of molecular and genetic testing, using samples obtained by R-EBUS, was satisfactory.

Efficacy and Radiation Exposure of Ultra-Low-Dose Chest CT at 100 kVp with Tin Filtration in CT-Guided Percutaneous Core Needle Biopsy for Small Pulmonary Lesions Using a Third-Generation Dual-Source CT Scanner

PurposeTo prospectively investigate efficacy and radiation dose of ultra-low-dose CT–guided percutaneous core needle biopsy (PCNB) at 100 kVp with tin filtration (100Sn kVp) for small pulmonary lesions. Materials and MethodsStudy enrolled and randomly assigned 210 patients to standard-dose CT (n = 70) or ultra-low-dose CT (n = 140; 1:2 randomization scheme) protocol. Standard-dose CT settings were reference 110 kVp and 50 mAs, and ultra-low-dose CT settings were fixed at 100Sn kVp and 70 mAs. All PCNBs in patients with small pulmonary lesions (< 3 cm) were performed on a third-generation dual-source CT scanner. Diagnostic performance, complication rate, image quality, and radiation dose were compared. ResultsSensitivity, specificity, and accuracy for diagnosis of malignancy were 95.7%, 100%, and 96.9% with standard-dose CT and 93.8%, 100%, and 95.4% with ultra-low-dose CT (P > .05). Complication rate showed no significant differences between protocols (P > .05). Mean volume CT dose index) and total dose-length product were significantly lower in ultra-low-dose CT compared with standard-dose CT (0.24 mGy vs 3.3 mGy ± 1.1 and 9.84 mGy-cm ± 0.70 vs 110.5 mGy-cm ± 45.1; P < .001). Effective dose for ultra-low-dose CT was significantly lower than that for standard-dose CT (0.14 mSv ± 0.02 vs 1.78 mSv ± 0.76; −92.1%; P < .001). Image quality of ultra-low-dose CT met the requirements of PCNB. ConclusionsUltra-low-dose CT-guided PCNB at 100Sn kVp spectral shaping significantly reduced radiation dose on a third-generation dual-source CT, while maintaining high diagnostic accuracy and safety for small pulmonary lesions.

Diagnostic yield of small histological cores obtained with a new EUS-guided fine needle biopsy system.

As endoscopic ultrasound-guided tissue acquisition techniques evolve, there is increasing interest in obtaining optimal histological samples to improve diagnostic accuracy. In this study, we aimed to assess the tissue acquisition success rate and test performance characteristics of a novel endoscopic ultrasound-guided fine needle biopsy (EUS-FNB) system. We performed a retrospective review of consecutive patients undergoing EUS-guided tissue sampling of solid lesions using the SharkCore fine needle system in a tertiary referral facility. At least two passes were submitted for histology and diagnostic accuracy was evaluated. Comparison standard was based on final surgical pathology or minimum six-month clinical follow-up. Seventy-nine patients underwent 85 EUS-FNB procedures. Of the 85 histology specimens, 78 (91.7%) were adequate for diagnostic examination (includes six atypical/suspicious for adenocarcinoma). The sensitivity, specificity, and accuracy for diagnosis of malignancy with FNB were 87.1, 100, and 90.6%, respectively. Cytology was simultaneously sent in 43 cases with the same needle in addition to histology. Out of the 14 cases that were atypical/suspicious for adenocarcinoma or non-diagnostic on cytology, 11 cases (78.6%) achieved definite diagnoses on histology. The overall sensitivity, specificity, and accuracy for diagnosis of malignancy combining histology and cytology were 90.3, 100, and 92.9%, respectively. No complications were reported after the procedures. In this initial experience with a new EUS-guided FNB system, obtaining small cores to submit for histological analysis was safe, technically feasible, and highly accurate. Most of the histological cores obtained via FNB yielded a definite diagnosis including in cases with equivocal cytomorphology. Further study is required to confirm these findings.

CT-guided transthoracic core needle biopsy for small pulmonary lesions: diagnostic performance and adequacy for molecular testing.

Computed tomography (CT)-guided transthoracic needle biopsy is a well-established, minimally invasive diagnostic tool for pulmonary lesions. Few large studies have been conducted on the diagnostic performance and adequacy for molecular testing of transthoracic core needle biopsy (TCNB) for small pulmonary lesions. This study included CT-guided TCNB with 18-gauge cutting needles in 560 consecutive patients with small (≤3 cm) pulmonary lesions from January 2012 to January 2015. There were 323 males and 237 females, aged 51.8±12.7 years. The size of the pulmonary lesions was 1.8±0.6 cm. The sensitivity, specificity, accuracy and complications of the biopsies were investigated. The risk factors of diagnostic failure were assessed using univariate and multivariate analyses. The sample's adequacy for molecular testing of non-small cell lung cancer (NSCLC) was analyzed. The overall sensitivity, specificity, and accuracy for diagnosis of malignancy were 92.0% (311/338), 98.6% (219/222), and 94.6% (530/560), respectively. The incidence of bleeding complications was 22.9% (128/560), and the incidence of pneumothorax was 10.4% (58/560). Logistic multivariate regression analysis showed that the independent risk factors for diagnostic failure were a lesion size ≤1 cm [odds ratio (OR), 3.95; P=0.007], lower lobe lesions (OR, 2.83; P=0.001), and pneumothorax (OR, 1.98; P=0.004). Genetic analysis was successfully performed on 95.45% (168/176) of specimens diagnosed as NSCLC. At least 96.8% of samples with two or more passes from a lesion were sufficient for molecular testing. The diagnostic yield of small pulmonary lesions by CT-guided TCNB is high, and the procedure is relatively safe. A lesion size ≤1 cm, lower lobe lesions, and pneumothorax are independent risk factors for biopsy diagnostic failure. TCNB specimens could provide adequate tissues for molecular testing.

Diagnostic value of tumour markers Ca-125 and CEA in the diagnostics of malignant pleural fluids

Abstract The significance of carbohydrate (cancer) antigen 125 (Ca-125) and carcinoembryonic antigen (CEA) tumour markers levels in differential diagnostics of malignant and benign pleural effusion was studied. Within this prospective study, 121 patients with fluids of various aetiology in the pleural cavity were analysed. Malignant pleural effusion was detected in 55 patients, parapneumonic effusion in 28 patients, transudative effusion of cardiac origin in 31 patients, pancreatitis in one patient and tuberculous pleurisy in five patients. The highest accuracy in diagnosis of malignancy was observed for Ca-125 and CEA levels in the pleural fluid: 75.2% at cut-off value ≥1452 U/mL and 76.9% at cut-off value ≥6.58 ng/mL, respectively. We conclude that the level of tumour markers in pleural fluid has additional diagnostic significance in the differential diagnosis of malignant and benign pleural effusions.

C-Arm Cone-Beam CT-guided Percutaneous Transthoracic Needle Biopsy of Lung Nodules: Clinical Experience in 1108 Patients

To retrospectively evaluate the diagnostic performance and complications of C-arm cone-beam computed tomography (CT)-guided percutaneous transthoracic needle biopsy (PTNB) in 1108 patients. This retrospective study was approved by the institutional review board with waiver of patient informed consent. From January 2009 to December 2011, 1108 patients (633 male, 475 female; mean age, 62.4 years ± 12.3 [standard deviation]) with 1116 pulmonary lesions (mean size, 2.7 cm ± 1.7) underwent 1153 cone-beam CT-guided PTNBs. A coaxial system with 18-gauge cutting needles was used. Diagnostic performance, complication rate, influencing factors, and patient radiation exposure were investigated. Variables influencing diagnostic performance and complications were assessed by using uni- and multivariate logistic regression analyses. Among 1153 PTNBs, pathologic analysis showed 1148 (99.6%) were technically successful (766 malignant [66.4%], 323 benign [28.0%], and 59 [5.1%] indeterminate). Sensitivity, specificity, and accuracy for diagnosis of malignancy were 95.7% (733 of 766), 100% (323 of 323), and 97.0% (1056 of 1089), respectively. In regard to diagnostic failures (five technical failures, 33 false-negative findings), lesions 1 cm in diameter or smaller and lesions in the lower lobe were significant risk factors (P = .028 and P = .034, respectively). As for complications, pneumothorax and hemoptysis occurred in 196 (17.0%) and 80 (6.9%) procedures, respectively. Multivariate analysis revealed two or more pleural passages and emphysema along the needle pathway were the two most significant risk factors for pneumothorax, and ground-glass nodules were the most significant risk factor for hemoptysis (P < .001 for all). Virtual guidance was a significant protective factor for both pneumothorax and hemoptysis (P < .001 for both). Mean estimated effective radiation dose through cone-beam CT-guided PTNBs was 7.3 mSv ± 4.1. Cone-beam CT-guided PTNB is a highly accurate and safe technique with which to diagnose pulmonary lesions with reasonable radiation exposure.

Feasibility and efficiency of a new 22G core needle: a prospective comparison study

Histological examination of core tissue samples may have advantages over cytology in endoscopic ultrasound (EUS)-guided sampling. We aimed to evaluate the feasibility and efficiency of a new 22G core biopsy needle. Consecutive patients with a pancreatic mass lesion or peri-intestinal lymphadenopathy sequentially underwent fine needle biopsy with both a newly developed 22G core needle (the FNB needle) and a standard 22G fine needle aspiration (FNA) needle, in randomized order. In 144 patients, mean age 48 years (± standard deviation [SD] 14; range 18 - 82), with 145 lesions (mean lesion size 39 ± 15 mm, range 15 - 99), EUS-guided sampling was technically feasible with both needles in all patients. Mean number of passes to obtain sufficient tissue was 1.2 ± 0.5 with the core needle vs. 2.5 ± 0.9 with the standard needle (P < 0.001). FNB specimens were adequate for evaluation in 125 (86.2 %) vs. 127 (87.6 %) with FNA (P = 0.72). Among 139 patients available for follow-up, FNB provided a correct diagnosis in 110 (79.1 %) and FNA in 112 (80.6 %) (P = 0.73). Sensitivity, specificity, positive and negative predictive values, and accuracy for diagnosis of malignancy were 90 %, 100 %, 100 %, 93 %, 96 % for FNB and 77 %, 100 %, 100 %, 85 %, 92 % for FNA, respectively (P > 0.05). FNB with the new 22G core needle was technically feasible, efficient and comparable to FNA with a standard needle. The core needle required fewer passes to provide an adequate sample, offering potentially shorter procedure time.

The reasons of false negative results of endobronchial ultrasound-guided transbronchial needle aspiration in the diagnosis of intrapulmonary and mediastinal malignancy.

Although the sensitivity can reach 99%, endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) has a significantly high false negative rate for diagnosis and staging of thoracic malignancy. We performed this retrospective study to investigate the causes of false negative results and to improve the efficacy and accuracy of EBUS-TBNA. We reviewed all patients suspected of intrapulmonary or mediastinal malignancy who undertook EBUS-TBNA between July 2009 and August 2012 in Chinese PLA general hospital. We retrieved the pathological results of EBUS-TBNA and video-assisted thoracic surgery (VATS) and follow-up data. The sensitivity, specificity, positive predictive value and negative predictive value were calculated. 185 patients were included in this study. Diagnosis of malignancy was established on 172 patients by EBUS-TBNA, and 8 patients with negative EBUS-TBNA result gained their final diagnosis of malignancy via VATS. The sensitivity, specificity, negative predictive value and accuracy for diagnosis of malignancy for EBUS-TBNA were 96%, 100%, 33% and 96% respectively. Inadequacy of the EBUS-TBNA specimens, internal necrosis in the lymph nodes and rare cancer types contributed to the false negative EBUS-TBNA results. VATS is obligatory to explain the negative results of EBUS-TBNA in patients suspected of malignancy.

Improving the classification accuracy for IR spectroscopic diagnosis of stomach and colon malignancy using non-linear spectral feature extraction methods

Non-linear feature extraction methods, neighborhood preserving embedding (NPE) and supervised NPE (SNPE), were employed to effectively represent the IR spectral features of stomach and colon biopsy tissues for classification, and improve the classification accuracy for diagnosis of malignancy. The motivation was to utilize the NPE and SNPE's capability of capturing non-linear spectral behaviors by simultaneously preserving local relationships in order that minute spectral differences among classes would be effectively recognized. NPE and SNPE derive an optimal embedding feature such that the local neighborhood structure can be preserved in reduced spaces (variables). The IR spectra collected from stomach and colon tissues were represented by several new variables through NPE and SNPE, and also by using the principal component analysis (PCA). Then, the feature-extracted variables were subsequently classified into normal, adenoma and cancer tissues by using both k-nearest neighbor (k-NN) and support vector machine (SVM), and the resulting accuracies were compared with each other. In both cases, the combination of SNPE-SVM provided the best classification performance, and the accuracy was substantially improved compared to when PCA-SVM was used. Overall results demonstrate that NPE and SNPE could be potential feature-representation strategies useful in biomedical diagnosis based on vibrational spectroscopy where effective recognition of minute spectral differences is critical.

Polysomy and p16 deletion by fluorescence in situ hybridization in the diagnosis of indeterminate biliary strictures

The diagnosis of indeterminate biliary strictures is limited because of the low sensitivity of cytology. However, an accurate diagnosis of malignancy is critical in the management of patients with suspected biliary malignancy. Testing for chromosomal aneuploidy by fluorescence in situ hybridization (FISH) may increase the yield. To evaluate the diagnostic accuracy of FISH in indeterminate biliary strictures and the additional value of including deletion of 9p21 (p16) in the diagnostic criteria of malignant biliary strictures. Retrospective review. Academic medical center. This study involved 76 consecutive patients who were seen for the evaluation of indeterminate strictures at our institution. These patients were screened, and 50 patients with either a final pathologic diagnosis or ≥ 12 months' conclusive follow-up were included in the analysis. Sensitivity, specificity, and area under the curve (AUC) analysis of cytology alone compared with the presence of FISH polysomy versus FISH polysomy and 9p21 deletion. The presence of increased copy numbers (polysomy) of chromosome 3, 7, or 17 by FISH increased the sensitivity of brush cytology from 21% to 58%, and when the presence of 9p21 deletion was included, the sensitivity increased to 89%. The specificity of FISH was 97% (vs 100% for cytology). The accuracy of cytology combined with FISH polysomy (AUC = 0.93) or p16 deletion was significantly greater than the accuracy of cytology alone (AUC 0.6; P < .001) or even cytology combined with FISH polysomy (AUC = 0.77; P ≤ .05). Sample size. There is a relatively high incidence of malignant biliary strictures in the entire cohort but low incidence among primary sclerosing cholangitis patients, and the majority of cancers are cholangiocarcinomas (as opposed to pancreatic). FISH significantly improves the diagnostic accuracy of brush cytology in indeterminate biliary strictures. In our series, the addition of 9p21 deletion to FISH polysomy and cytology further improved sensitivity. This suggests that 9p21 deletion may be added to the diagnostic criteria in indeterminate strictures.