Number Of False-positive Findings Research Articles

Abstract P5-01-09: Improving specificity and refining diagnostic accuracy of MRI in breast cancer with dedicated breast PET (dbPET)

Abstract BACKGROUND. Continued progress in the control of breast cancer will require sustained efforts to provide high-quality screening, diagnosis, and treatment to all segments of the population. MRI has the advantages of providing a 3D view of the breast with high sensitivity, using non-ionizing radiation. However, MRI has significant limitations including its moderate specificity that, in combination with high sensitivity, often leads to unnecessary biopsies. Recently, the MAMmography with Molecular Imaging (MAMMI) dedicated breast PET (dbPET) has emerged as an additional imaging tool for breast cancer diagnosis, clarification of complex lesions and therapy follow-up. It is well known that 18F-FDG-PET has high specificity by assessing metabolic activity, potentially reducing the number of false positive findings. To compare FDG-dbPET with the conventional magnetic resonance imaging (MRI) on breast lesion characterization, we analyzed sensitivity, specificity, Positive Predictive Value (PPV) and Negative Predictive Value (NPV) of 37 cases of patients with BC using both dbPET and MRI. METHODS. Thirty-seven women with known or suspected breast carcinoma (41 lesions: 36 invasive carcinomas, 2 noninvasive carcinomas, 1 case of ductal hyperplasia, and 2 benign lesions) were enrolled in this study. Both a prone dbPET Mammi scan, and routine breast MRI scans were performed. A joint reading of MRI and PET scans side-by-side by a nuclear medicine physician and a radiologist was performed. Sensitivity and specificity of MRI and dbPET scans were calculated on the basis of post-surgical pathology reports. Breast MRI examinations were performed in a 1.5-T or 3-T commercial imager (Siemens Medical Solutions) with the use of a dedicated breast coil. The imaging sequence included a sagittal T1-weighted localizing sequence followed by a sagittal T2- weighted sequence. A T1-weighted 3D, fat-suppressed fast spoiled-gradient-echo sequence was then performed with an injection of 0.1 mmol per kilogram body weight of gadolinium dimeglumine (Magnevist; Schering). A prone position high-resolution dedicated breast PET was performed 60 min after administration of 90-120 MBq of 18F-FDG. RESULTS. A total of 41 lesions were assessed. Lesion size range was 0.2 to 7.6 cm. In lesion-by-lesion analysis, sensitivity and specificity of MRI alone were 91% and 54%, respectively; while lesion-based sensitivity of dbPET was 93% and breast-based specificity was 100%. The positive predictive value and the negative predictive value for MRI alone were 69% and 85%, respectively; and for dbPET were 100% and 89%, respectively. In a significant number of cases, dbPET helped to clarify or disprove positive findings by MRI, and in four cases helped to define new positives that had gone unnoticed at MRI, CONCLUSION: Dedicated breast PET scans increase the specificity of MRI. False positives, one of the most challenging aspects of MRI in breast lesions, are reduced. The results of the current study show that FDG-dbPET is more effective than MRI in detecting true breast cancer positives. Its functional information may improve the likelihood of a successful excision, reduce costs from additional procedures, and minimize discomfort and anxiety for the patient. Citation Format: Ines Dominguez, Michel Herranz, Sze Yiun Teo, Elena Brozos, Carmela Rodriguez, Jasper Chaal, Juan Cueva, Jose Ramon Antúnez, Gabriel Gonzalez Pavón, Marielle Fortier, Rafael López, Alvaro Ruibal. Improving specificity and refining diagnostic accuracy of MRI in breast cancer with dedicated breast PET (dbPET) [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P5-01-09.

Lung cancer screening-don't forget the chest radiograph.

Lung cancer is a major health burden and early detection only bears the possibility of curative treatment. Screening with computed tomography (CT) recently demonstrated a mortality reduction in selected patients and has been incorporated in clinical guidelines. Problems of screening with CT are the excessive number of false positive findings, costs, radiation burden and from a global point of view shortage of CT capacity. In contrast, chest radiography could be an ideal screening tool in the early detection of lung cancer. It is widely available, easy to perform, cheap, the radiation burden is negligible and there is only a low rate of false positive findings. Large randomized controlled trials could not show a mortality reduction, but different large population-based cohort studies have shown a lung cancer mortality reduction. It has been argued that community-based cohort studies are more closely reflecting the "real world" of everyday medicine. Radiologists should be aware of the found mortality reduction and realize that early detection of lung cancer is possible when reading their daily chest radiographs. Offering a chest radiograph in selected scenarios for the early detection of lung cancer is therefore still justified.

Radiolabelled leucocyte scintigraphy versus conventional radiological imaging for the management of late, low-grade vascular prosthesis infections

In this study we evaluated the diagnostic performance of (99m)Tc-HMPAO-leucocyte ((99m)Tc-HMPAO-WBC) scintigraphy in a consecutive series of 55 patients (46 men and 9 women, mean age 71 ± 9 years, range 50 - 88 years) with a suspected late or a low-grade late vascular prosthesis infection (VPI), also comparing the diagnostic accuracy of WBC with that of other radiological imaging methods. All patients suspected of having VPI underwent clinical examination, blood tests, microbiology, US and CT, and were classified according to the Fitzgerald criteria. A final diagnosis of VPI was established in 47 of the 55 patients, with microbiological confirmation after surgical removal of the prosthesis in 36 of the 47. In the 11 patients with major contraindications to surgery, the final diagnosis was based on microbiology and clinical follow-up of at least 18 months. (99m)Tc-HMPAO-WBC planar, SPECT and SPECT/CT imaging identified VPI in 43 of 47 patients (20 of these also showed infection at extra-prosthetic sites). In the remaining eight patients without VPI, different sites of infections were found. The use of SPECT/CT images led to a significant reduction in the number of false-positive findings in 37% of patients (sensitivity and specificity 100 %, versus 85.1% and 62.5% for stand-alone SPECT). Sensitivity and specificity were 34% and 75% for US, 48.9% and 83.3% for CT, and 68.1% and 62.5% for the FitzGerald classification. Perioperative mortality was 5.5%, mid-term mortality 12%, and long-term mortality 27%. Survival rates were similar in patients treated with surgery and antimicrobial therapy compared to patients treated with antimicrobial therapy alone (61% versus 63%, respectively), while infection eradication at 12 months was significantly higher following surgery (83.3% versus 45.5%). (99m)Tc-HMPAO-WBC SPECT/CT is useful for detecting, localizing and defining the extent of graft infection in patients with late and low-grade late VPI with inconclusive radiological findings. (99m)Tc-HMPAO-WBC SPECT/CT might be used to optimize patient management.

Clinical utility of CT and FDG PET/CT in assessing the neck in node-positive head and neck cancer after chemoradiotherapy.

6082 Background: Concurrent chemoradiotherapy has been widely accepted to treat locoregional advanced head and neck cancer, but the need for subsequent neck dissection remains controversial. Our objective was to determine whether CT or fluorodeoxyglucose (FDG) PET/CT is superior in the evaluation of persistent nodal disease after chemoradiotherapy in patients with node-positive head and neck squamous cell carcinoma (HNSCC). Methods: Study entry criteria included node-positive HNSCC treated with concurrent chemoradiotherapy, a local complete response, and post-treatment CT and FDG PET/CT studies 11 weeks after chemoradiotherapy. Fifty-eight patients with 68 node-positive necks were eligible. Nodes larger than 1 cm (minor axis), or with central necrosis on CT, or any visually hypermetabolic nodes on FDG PET/CT were considered clinically positive. Regardless of PET/CT findings, necks with positive CT were subjected to neck dissection, whereas those with negative CT were observed without neck dissection. Results: Seventeen necks showed positive CT, 13 and 4 of which underwent neck dissection and fine needle aspiration cytology, respectively, resulting in pathologic evidence of persistent nodal disease in 5 necks. Four of 51 necks with negative CT developed regional recurrence. Diagnostic accuracy of CT and PET/CT is shown in table. In general, the negative predictive value (NPV) was equivalent between CT and FDG PET/CT, whereas FDG PET/CT was better than CT in the specificity and accuracy. Conclusions: In patients with HNSCC, both CT and FDG PET/CT after chemoradiotherapy have a high NPV for excluding residual regional disease and avoiding unnecessary neck dissection. Although the NPV is similarly high, PET/CT has superior utility compared with CT because the number of false positive findings is less in PET/CT than CT. [Table: see text]

Automatic detection of lytic and blastic thoracolumbar spine metastases on computed tomography

ObjectiveTo evaluate a computer-aided detection (CADe) system for lytic and blastic spinal metastases on computed tomography (CT).MethodsWe retrospectively evaluated the CADe system on 20 consecutive patients with 42 lytic and on 30 consecutive patients with 172 blastic metastases. The CADe system was trained using CT images of 114 subjects with 102 lytic and 308 blastic spinal metastases. Lesions were annotated by experienced radiologists. Detected benign lesions were considered false-positive findings. Detector sensitivity and the number of false-positive findings were calculated as the criteria for detector performance, and free-response receiver operating characteristic (FROC) analysis was conducted. Detailed analysis of false-positive and false-negative findings was performed.ResultsAlgorithm runtime is 3 ± 0.5 min per patient. The system achieves a sensitivity of 83 % at 3.5 false positives per patient on average for blastic metastases and a sensitivity of 88 % at 3.7 false positives for lytic metastases. False positives appeared predominantly in the area of degenerative changes in the case of the blastic metastasis detector and in osteoporotic areas in the case of the lytic metastasis detector.ConclusionThe CADe system reliably detects thoracolumbar spine metastases in real time. An additional study is planned to evaluate how the bone lesion CADe system improves radiologists’ accuracy and efficiency in a clinical setting.Key Points• Computer-aided detection (CADe) of bone metastases has been developed for spinal CT.• The CADe system exhibits high sensitivity with a tolerable false-positive rate.• Analysis of false-positive detection may further improve the system.• CADe may reduce the number of missed spinal metastases at CT interpretation.

Post-run target screening strategy for ultra high performance liquid chromatography coupled to Orbitrap based veterinary drug residue analysis in animal urine

The performance of liquid chromatography coupled to high resolution mass spectrometry (LC–HRMS) post-run target screening for veterinary drug residue analysis (sulfonamides, tetracyclines and quinolones) in animal urine has been critically evaluated. It was found that retention time information still remains an essential information and that accurate masses together with relative isotopic abundance data alone are not sufficient for many residue applications. Post-run target screening requires the careful setting of parameters to achieve near zero false negative (above a defined threshold level) and a manageable numbers of false positive findings. HRMS offers many possibilities for the reduction of false positives (e.g. isotopic ratio, isotopic fine structure, exact mass of fragment ions). However, the successful use of such tools requires a sufficient ion intensity. This is often not available when trace level compounds are to be detected. Nevertheless, the proposed method is sufficiently sensitive to detect the veterinary drugs at the relevant concentration levels in urine. This means that the approach is well suited to significantly reduce the number of corresponding meat samples which have to be analyzed in a final step for the regulatory relevant quantification of residue levels in meat. The semi-quantitative screening of many samples for a large number of analytes within a short period of time requires the availability of software tools which provide fast and reliable answers.

Bayesian Multiple Quantitative Trait Loci Mapping for Recombinant Inbred Intercrosses

The Collaborative Cross (CC) is a renewable mouse resource that mimics the genetic diversity in humans. The recombinant inbred intercrosses (RIX) generated from CC recombinant inbred (RI) lines share similar genetic structures to those of F(2) individuals. In contrast to F(2) mice, genotypes of RIX can be inferred from the genotypes of their RI parents and can be produced repeatedly. Also, RIX mice do not typically share the same degree of relatedness. This unbalanced genetic relatedness requires careful statistical modeling to avoid a large number of false positive findings. For complex traits, mapping multiple genes simultaneously is arguably more powerful than mapping one gene at a time. In this article, we describe how we have developed a Bayesian quantitative trait locus (QTL) mapping method that simultaneously deals with the special genetic architecture of RIX and maps multiple genes. The performance of the proposed method is evaluated by extensive simulations. In addition, for a given set of RI lines, there are numerous ways to generate RIX samples. To provide a general guideline on future RIX studies, we compare several RIX designs through simulations.

Clinical MR Mammography: Impact of Hormonal Status on Background Enhancement and Diagnostic Accuracy

Hormonal stimulation can induce background enhancement (BE) in MR mammography (MRM). This fact has been assumed to decrease the accuracy of MRM. Consequently, this report investigates: 1. The prevalence of BE in postmenopausal vs. premenopausal women in correlation to hormonal cycle phase (CP). 2. The impact of hormonal status (HS) and BE on diagnostic accuracy. Consecutive patients over 22 months with complete HS information (week of CP or postmenopausal) were included in this prospective investigation. Exclusion criteria were any hormonal therapy, hysterectomy as well as cancer proven by biopsy. The standard of reference was histopathology. All MRM scans were acquired using the same protocol (1.5 T, dynamic T 1w GRE after 0.1 mmol/kg bw Gd-DTPA i. v.). Two radiologists rated all examinations in consensus according to BI-RADS. BE was defined as: 0 = missing, 1 = moderate, 2 = distinct. 224 patients (150 postmenopausal, 74 premenopausal, 45 in the second week of CP) were included in this study (83 benign and 141 malignant findings). BE was more frequent in premenopausal women (p = 0.006), but did not differ between CP (p = 0.460). Neither HS nor BE had a significant impact on the diagnostic parameters of MRM (p ≥ 0.375). However, regarding BE, the relative number of false positive (FP) findings was highest (5 / 10; 50 %) in the distinct BE group. Regarding HS, 17 % more FP findings were observed in premenopausal women examined outside the second week of CP. In premenopausal women, HS leads to increased BE of breast tissue, independent of CP. Distinct BE and less pronounced, non-optimal CP may lead to an increased number of false positive findings.

Impact of Image Quality on the Performance of Computer-Aided Detection of Pulmonary Embolism

The purpose of this article is to assess the relationship between CT image quality and the number and type of false-positive (FP) findings found by a prototype computer-aided detection (CAD) algorithm for automatic detection of pulmonary embolism (PE). This retrospective study included 278 subjects (138 men and 140 women; mean age, 57 years; range, 18-88 years) who underwent consecutive CT pulmonary angiographies performed during off hours. Twenty-four percent (68/278) of studies were reported as positive for PE. CAD findings were classified as true-positive or FP by two independent readers and, in cases of discordance, by a third radiologist. Each FP result was classified according to underlying cause. The degree of vascular enhancement, image noise, motion artifacts, overall quality, and presence of underlying lung disease were rated on a 4- or 5-point scale. Chi-square tests and t tests were used to test significance of differences. The mean number of FP CAD findings was 4.7 (median, 2) per examination. Most were caused by veins (30% [389/1,298]) or airspace consolidations (22% [286/1,298]). There was a significant positive association between the number of FP findings and image noise, motion artifacts, low vascular enhancement, low overall quality, and the extent of underlying disease. On a per-embolism basis, sensitivity decreased from 70.6% (214/303) for scans with zero to five FP findings, to 62.3% (33/53) for scans with six to 10 FP findings, to 60% (12/20) for scans with more than 10 FP findings. There is a strong association between CT image quality and the number of FP findings indicated by a CAD algorithm for the detection of PE.

High signal in bone marrow at diffusion-weighted imaging with body background suppression (DWIBS) in healthy children

In our experience, diffusion-weighted imaging with body background suppression (DWIBS) is hard to interpret in children who commonly have foci of restricted diffusion in their skeletons unrelated to pathology, sometimes in an asymmetrical pattern. This raises serious concern about the accuracy of DWIBS in cancer staging in children. To describe the signal distribution at DWIBS in the normal developing lumbar spine and pelvic skeleton. Forty-two healthy children underwent an MR DWIBS sequence of the abdomen and pelvis. An axial short-tau inversion-recovery (STIR) echo-planar imaging (EPI) pulse sequence was used. Two radiologists did a primary review of the images and based on these preliminary observations, separate scoring systems for the lumbar spine, pelvis and proximal femoral epiphyses/femoral heads were devised. Visual evaluation of the images was then performed by the two radiologists in consensus. The scoring was repeated separately 2months later by a third radiologist. Restricted diffusion was defined as areas of high signal compared to the background. Coronal maximum intensity projection (MIP) reformats were used to assess the vertebral bodies. For the pelvis, the extension of high signal for each bone was given a score of 0 to 4. Cohen's Kappa interobserver agreement coefficients of signal distribution and asymmetry were calculated. All children had areas of high signal, both within the lumbar vertebral bodies and within the pelvic skeleton. Three patterns of signal distribution were seen in the lumbar spine, but no specific pattern was seen in the pelvis. There was a tendency toward a reduction of relative area of high signal within each bone with age, but also a widespread interindividual variation. Restricted diffusion is a normal finding in the pelvic skeleton and lumbar spine in children with an asymmetrical distribution seen in 48% of normal children in this study. DWIBS should be used with caution for cancer staging in children as this could lead to high numbers of false positive findings or even unjustified upstaging.

How useful is the Social Communication Questionnaire in toddlers at risk of autism spectrum disorder?

The Social Communication Questionnaire (SCQ) is a screening instrument with established validity against the Autism Diagnostic Interview-Revised (ADI-R) in children aged 4 years and older. Indices of diagnostic accuracy have been shown to be strong in school-aged samples; however, relatively little is known about the performance of the SCQ in toddlers at risk of autism spectrum disorder (ASD). This study replicates and extends previous research by Corsello et al. (2007) in a comparatively large (N = 208), substantially younger (20-40 months) sample of children at high risk of ASD. The usefulness of the SCQ as a second-level screening instrument with different cut-off scores was evaluated in relation to IQ, age, and type of ASD diagnosis. The use of the SCQ as compared to the ADI-R was evaluated against clinical diagnosis, both alone and in combination with the ADOS. The SCQ with different cut-offs consistently showed an unsatisfactory balance between sensitivity and specificity in screening for ASD in high-risk toddlers, with only a few exceptions for specific age, IQ, or diagnostic groups. Even though the SCQ and ADI-R were highly correlated, diagnostic agreement with the best evidence clinical diagnosis was poor for both measures. The ADOS used alone consistently had the highest predictive value. For autism versus not-autism, the combined SCQ and ADOS performed as well as the ADOS alone and notably better than the combination ADI-R and ADOS. The SCQ is likely to result in a number of false-positive findings, particularly in children with autism symptomatology, and the balance between sensitivity and specificity is poor. The ADOS should be considered the most valid and reliable diagnostic instrument in these very young at-risk children.

PERMORY: an LD-exploiting permutation test algorithm for powerful genome-wide association testing

In genome-wide association studies (GWAS) examining hundreds of thousands of genetic markers, the potentially high number of false positive findings requires statistical correction for multiple testing. Permutation tests are considered the gold standard for multiple testing correction in GWAS, because they simultaneously provide unbiased type I error control and high power. At the same time, they demand heavy computational effort, especially with large-scale datasets of modern GWAS. In recent years, the computational problem has been circumvented by using approximations to permutation tests, which, however, may be biased. We have tackled the original computational problem of permutation testing in GWAS and herein present a permutation test algorithm one or more orders of magnitude faster than existing implementations, which enables efficient permutation testing on a genome-wide scale. Our algorithm does not rely on any kind of approximation and hence produces unbiased results identical to a standard permutation test. A noteworthy feature of our algorithm is a particularly effective performance when analyzing high-density marker sets. Freely available on the web at http://www.permory.org.

Nationwide skin cancer screening in Germany

Skin cancer is the most frequent form of cancer. Altogether, every year in Germany more than 195,000 people are newly diagnosed as having skin cancer (basal cell, prickle cell, malignant melanoma carcinoma). The joint federal committee decided on the nationwide introduction of skin cancer screening from July 1st, 2008. Their decision was based on the results found in Schleswig-Holstein. From the age of 35 years, everyone has the right to an early diagnosis examination every two years. The worldwide largest quality assured skin cancer screening programme began in Germany. The objective of early screening for skin cancer is to discover malignant melanoma, basal cell carcinoma and spinocellular carcinoma in time. The examination includes targeted medical history, visual, according to the education program standardized full body inspection of the entire skin including the scalp and all intertriginous areas, findings report with related consultation and documentation. If the visual inspection of the skin by a general practitioner gives rise to suspicion of the presence of a target disease, then additional clarification by a specialist for skin and venereal diseases is to be performed. This physician carries out an additional full-body inspection, if he is not the first examiner, he checks in particular abnormal findings of the previous examiner and, where applicable, orders the histopathological examination to confirm the diagnosis. The histopathological evaluation can only be performed by pathologists as well as by dermatologists with additional credentials in dermatohistology corresponding to the stipulations of the respective further education ordinance. Obligatory prerequisite for approval by the responsible SHI-accredited physician association regarding the implementation and billing of early screening for skin cancer is proof of successful completion of an eight-hour further education program certified by the SHI-accredited physician association. The course must include the following subjects in particular: (1) potential benefits and risks of early screening measures, criteria for evaluating early screening measures (2) program for early screening, health examinations and early awareness of the patient (3) measures regarding how to address insurance holders (4) etiology of skin cancer, clinical pattern, frequency, risk factors or groups, medical history, standardized visual full-body inspection, visual diagnostics (5) process of early screening for skin cancer (6) presentation and discussion of case examples (7) documentation measures (8) interdisciplinary cooperation. The measures for ‘early screening of skin cancer’ are evaluated with regard to quality and goal achievement through regular evaluation of the documentation. During this process the National Association of Statutory Health Insurance Physicians and the Head Association of Health Insurance Funds agree on the type and scope of evaluation and publication. The objective parameters for the evaluation include in particular the following: (1) participant rates differentiated according to physician groups, age, sex (2) simultaneous utilization of the health screening by family doctors (3) number of suspected diagnoses differentiated according to physician groups (4) number of diagnoses confirmed by dermatologists (5) number of false-positive findings (6) detection rate (participant rate/number of discovered skin cancers and the histopathological grading) (7) evaluation differentiated according to the SHI-accredited physician association areas. To answer additional specific questions separate studies (e.g. for determining the number of false-negative findings, moving the diagnosis time forward) should be carried out. If necessary, the data from the routine documentation will be made available for this while adhering to the data protection guidelines.

A Robust Procedure for Comparing Multiple Means under Heteroscedasticity in Unbalanced Designs

Investigating differences between means of more than two groups or experimental conditions is a routine research question addressed in biology. In order to assess differences statistically, multiple comparison procedures are applied. The most prominent procedures of this type, the Dunnett and Tukey-Kramer test, control the probability of reporting at least one false positive result when the data are normally distributed and when the sample sizes and variances do not differ between groups. All three assumptions are non-realistic in biological research and any violation leads to an increased number of reported false positive results. Based on a general statistical framework for simultaneous inference and robust covariance estimators we propose a new statistical multiple comparison procedure for assessing multiple means. In contrast to the Dunnett or Tukey-Kramer tests, no assumptions regarding the distribution, sample sizes or variance homogeneity are necessary. The performance of the new procedure is assessed by means of its familywise error rate and power under different distributions. The practical merits are demonstrated by a reanalysis of fatty acid phenotypes of the bacterium Bacillus simplex from the “Evolution Canyons” I and II in Israel. The simulation results show that even under severely varying variances, the procedure controls the number of false positive findings very well. Thus, the here presented procedure works well under biologically realistic scenarios of unbalanced group sizes, non-normality and heteroscedasticity.

Additional value of integrated PET/CT over PET alone in the initial staging and follow up of head and neck malignancy

Clinical application of FDG-PET in head and neck cancer includes identification of metastases, unknown primary head and neck malignancy, or second primary carcinoma, and also recurrent tumor after treatment. In this study, the additional value of PET/CT fusion images over PET images alone was evaluated in patients with initial staging and follow up of head and neck malignancy. Forty patients with suspected primary head and neck malignancy and 129 patients with suspected relapse after treatment of head and neck malignancy were included. FDG-PET/CT study was performed after the intravenous administration of FDG (5 MBq/kg). Target of evaluation was set at primary tumor, cervical lymph node, and whole body. PET images and PET with CT fusion images were compared. Sensitivity, specificity, accuracy, positive predictive value (PPV), and negative predictive value (NPV) were calculated. Results of PET and PET/CT were compared with postoperative histopathological examination, and case by case comparison of PET and PET/CT results for each region was performed. The additional value of CT images over PET only images was assessed. Statistical differences in sensitivity and specificity were evaluated. In the comparative evaluation of 507 targets by PET alone and PET/CT, 401 targets showed agreement of the results. Of the 106 discordant targets, 103 showed a positive result on PET alone and negative result on PET/CT. These results showed a significant difference (p< 0.01). Sensitivity of PET/CT was slightly higher than that of PET without statistical significance, while specificity of PET/CT was significantly higher than that of PET alone (Initial staging: 90.5% vs. 62.2%, p < 0.01; Follow up: 97.2% vs. 74.4%, p < 0.01). In Fisher's direct probability test, a significant difference was noted in the sensitivity (Initial staging: 91.3% vs. 87.0%, p < 0.01; Follow up: 93.9% vs. 91.4%, p <0.01). Combined PET/CT showed improved diagnostic performance than PET alone by decreasing the number of false positive findings in patients with initial staging and follow up of head and neck malignancy.

Identifying main effects and epistatic interactions from large-scale SNP data via adaptive group Lasso

BackgroundSingle nucleotide polymorphism (SNP) based association studies aim at identifying SNPs associated with phenotypes, for example, complex diseases. The associated SNPs may influence the disease risk individually (main effects) or behave jointly (epistatic interactions). For the analysis of high throughput data, the main difficulty is that the number of SNPs far exceeds the number of samples. This difficulty is amplified when identifying interactions.ResultsIn this paper, we propose an Adaptive Group Lasso (AGL) model for large-scale association studies. Our model enables us to analyze SNPs and their interactions simultaneously. We achieve this by introducing a sparsity constraint in our model based on the fact that only a small fraction of SNPs is disease-associated. In order to reduce the number of false positive findings, we develop an adaptive reweighting scheme to enhance sparsity. In addition, our method treats SNPs and their interactions as factors, and identifies them in a grouped manner. Thus, it is flexible to analyze various disease models, especially for interaction detection. However, due to the intensive computation when millions of interaction terms needs to be searched in the model fitting, our method needs to combined with some filtering methods when applied to genome-wide data for detecting interactions.ConclusionBy using a wide range of simulated datasets and a real dataset from WTCCC, we demonstrate the advantages of our method.

Polyp Characteristics Correctly Annotated by Computer-aided Detection Software but Ignored by Reporting Radiologists during CT Colonography

To retrospectively describe the characteristics of polyps incorrectly dismissed by radiologists despite appropriate computer-aided detection (CAD) prompting during computed tomographic (CT) colonography. Ethics committee approval and patient informed consent were obtained from institutions that provided the data sets used in this HIPAA-compliant study. A total of 111 polyps that had a diameter of at least 6 mm and were detected with CAD were collated from three previous studies in which researchers investigated radiologist performance with and without CAD (total, 25 readers). Two new observers graded each polyp with predefined criteria, including polyp size, morphology, and location; data set quality; ease of visualization; tagging use and polyp coating; colonic curvature; CAD mark obscuration; and number of false-positive findings. The 86 polyps that were missed before CAD (those that were unreported by one or more original readers) were divided into those that remained unreported after CAD (no CAD gain, n = 36) and those that were reported correctly by at least one additional reader (CAD gain, n = 50). Logistic-regression analysis and the Fisher exact and Mann-Whitney tests were used to compare the results of both groups with each other and with a control group of 25 polyps, all of which were detected by readers without CAD. Before CAD, polyps 10 mm in diameter or larger, those that were rated easy to visualize, and those that were uncoated by tagged fluid were less likely to be missed (72%, 76%, and 80% of control polyps vs 43%, 43%, and 59% of missed polyps, respectively; P < .001, P < .01, and P < .03, respectively). After CAD, the odds of CAD gain decreased with increasing polyp size (odds ratio, 0.92; 95% confidence interval: 0.85, 1.00; P = .04) and irregular morphology (odds ratio, 0.28; 95% confidence interval: 0.08, 0.92; P = .04). Larger irregular polyps are a common source of incorrect radiologist dismissal, despite correct CAD prompting.

Investigation of Reading Mode and Relative Sensitivity as Factors That Influence Reader Performance When Using Computer-Aided Detection Software

The aim of this study was to investigate the effects of relative sensitivity (reader without computer-aided detection [CAD] vs stand-alone CAD) and reading mode on reader performance when using CAD software. Two sets of 100 images (low-contrast and high-contrast sets) were created by adding low-contrast or high-contrast simulated masses to random locations in 100 normal mammograms. This produced a relative sensitivity, substantially less for the low-contrast set and similar for the high-contrast set. Seven readers reviewed every image in each set and specified location and probability scores using three reading modes (without CAD, second read with CAD, and concurrent read with CAD). Reader detection accuracy was analyzed using areas under free-response receiver operating characteristic curves, sensitivity, and the number of false-positive findings per image. For the low-contrast set, average differences in areas under free-response receiver operating characteristic curves, sensitivity, and false-positive findings per image without CAD were 0.02, 0.12, and 0.11, respectively, compared to second read and 0.05, 0.17, and 0.09 (not statistically significant), respectively, compared to concurrent read. For the high-contrast set, average differences were 0.002 (not statistically significant), 0.04, and 0.05, respectively, compared to second read and -0.004 (not statistically significant), 0.04, and 0.08 (not statistically significant), respectively, compared to concurrent read (all differences were statistically significant except as noted). Differences were greater in the low-contrast set than the high-contrast set. Differences between second read and concurrent read were not significant. Relative sensitivity is a critical factor that determines incremental improvement in reader performance when using CAD and appears to be more important than reading mode. Relative sensitivity may determine the clinical usefulness of CAD in different clinical applications and for different types of users.

Consensus Review of Discordant Findings Maximizes Cancer Detection Rate in Double-Reader Screening Mammography: Irish National Breast Screening Program Experience

To assesses consensus review of discordant screening mammography findings in terms of its sensitivity, safety, and effect on overall performance in the first 6 years of operation of the Irish National Breast Screening Program (NBSP). Women who participated in the Irish NBSP gave written informed consent for use of their data for auditing purposes. Local ethics committee approval was obtained. The study population consisted of women who participated in the Irish NBSP and underwent initial screening mammography at one of the two screening centers serving the eastern part of Ireland between 2000 and 2005. Independent double reading of mammograms was performed. When the readers disagreed regarding referral, the case was reviewed by a consensus panel. Of the 128 569 screenings performed, 1335 (1%) were discussed by consensus. Of the 1335 cases discussed by consensus, 606 (45.39%) were recalled for further assessment. This resulted in an overall recall rate of 4.41%. In those recalled to assessment, 71 cases of malignant disease were diagnosed (ductal carcinoma in situ, n = 24; invasive cancer, n = 47). The remaining 729 patients were returned to biennial screening. Of these 729 patients, seven had false-negative findings that were identified in the subsequent screening round. Use of the highest reader recall method, in which a patient is recalled if her findings are deemed abnormal by either reader, could potentially increase the cancer detection rate by 0.6 per 1000 women screened but would increase the recall rate by 12.69% and the number of false-positive findings by 15.37%. The consensus panel identified 71 (7.33%) of 968 cancers diagnosed. Consensus review substantially reduced the number of cases recalled and was associated with a low false-negative rate.

Visual evoked potentials in normal subjects and patients with multiple sclerosis.

Visual evoked potentials (VEPs) by checkerboard pattern-reversal stimulation were recorded in 70 subjects aged 10--69 years and in 100 patients with definite, probable or possible multiple sclerosis (MS). Longer latencies and smaller amplitudes of the major positive component were found in male subjects, in old subjects and when the amplifier's band-pass was narrowed. Subjects 10--14 years old had longer latencies and higher amplitudes than mature adults. Based on findings in the normal material, the following three criteria were used in evaluating the recordings from patients: the latency, the side difference in latencies and the ratio of amplitudes between the two sides of the major positive component with various limits for the two sexes and different age groups. The incidence of abnormal recordings was 85% for all the patients, 100% in 50 patients with definite, 70% in 50 patients with probable or possible MS, 73% in patients who had a history of spinal symptoms only, 98% if they had and 74% if they had not experienced optic neuritis. The incidence of abnormal findings increased with increasing duration of symptoms. All patients with visual acuity below 0.67 had abnormal VEPs. The high incidence of abnormal recordings confirmed the value of the test in establishing the diagnosis, and suggested that the use of different normative values for sex and age may increase the diagnostic yield without increasing the number of false positive findings.