Petrina A Causer

University of Toronto, Toronto, Ontario, Canada

Are you Petrina A Causer?

Claim your profile

Publications (45)

  • [Show abstract] [Hide abstract] ABSTRACT: Magnetic resonance imaging (MRI)-enabled cancer screening has been shown to be a highly sensitive method for the early detection of breast cancer. Computer-aided detection systems have the potential to improve the screening process by standardizing radiologists to a high level of diagnostic accuracy. This retrospective study was approved by the institutional review board of Sunnybrook Health Sciences Centre. This study compares the performance of a proposed method for computer-aided detection (based on the second-order spatial derivative of the relative signal intensity) with the signal enhancement ratio (SER) on MRI-based breast screening examinations. Comparison is performed using receiver operating characteristic (ROC) curve analysis as well as free-response receiver operating characteristic (FROC) curve analysis. A modified computer-aided detection system combining the proposed approach with the SER method is also presented. The proposed method provides improvements in the rates of false positive markings over the SER method in the detection of breast cancer (as assessed by FROC analysis). The modified computer-aided detection system that incorporates both the proposed method and the SER method yields ROC results equal to that produced by SER while simultaneously providing improvements over the SER method in terms of false positives per noncancerous exam. The proposed method for identifying malignancies outperforms the SER method in terms of false positives on a challenging dataset containing many small lesions and may play a useful role in breast cancer screening by MRI as part of a computer-aided detection system.
    Article · Aug 2015 · Journal of Digital Imaging
  • Jacob Levman · Ellen Warner · Petrina Causer · Anne Martel
    [Show abstract] [Hide abstract] ABSTRACT: Cancer screening with magnetic resonance imaging (MRI) is currently recommended for very high risk women. The high variability in the diagnostic accuracy of radiologists analyzing screening MRI examinations of the breast is due, at least in part, to the large amounts of data acquired. This has motivated substantial research towards the development of computer-aided diagnosis (CAD) systems for breast MRI which can assist in the diagnostic process by acting as a second reader of the examinations. This retrospective study was performed on 184 benign and 49 malignant lesions detected in a prospective MRI screening study of high risk women at Sunnybrook Health Sciences Centre. A method for performing semi-automatic lesion segmentation based on a supervised learning formulation was compared with the enhancement threshold based segmentation method in the context of a computer-aided diagnostic system. The results demonstrate that the proposed method can assist in providing increased separation between malignant and radiologically suspicious benign lesions. Separation between malignant and benign lesions based on margin measures improved from a receiver operating characteristic (ROC) curve area of 0.63 to 0.73 when the proposed segmentation method was compared with the enhancement threshold, representing a statistically significant improvement. Separation between malignant and benign lesions based on dynamic measures improved from a ROC curve area of 0.75 to 0.79 when the proposed segmentation method was compared to the enhancement threshold, also representing a statistically significant improvement. The proposed method has potential as a component of a computer-aided diagnostic system.
    Article · Aug 2014 · Journal of Digital Imaging
  • Jacob E D Levman · Ellen Warner · Petrina Causer · Anne L Martel
    [Show abstract] [Hide abstract] ABSTRACT: This study investigates the use of a proposed vector machine formulation with application to dynamic contrast-enhanced magnetic resonance imaging examinations in the context of the computer-aided diagnosis of breast cancer. This paper describes a method for generating feature measurements that characterize a lesion's vascular heterogeneity as well as a supervised learning formulation that represents an improvement over the conventional support vector machine in this application. Spatially varying signal-intensity measures were extracted from the examinations using principal components analysis and the machine learning technique known as the support vector machine (SVM) was used to classify the results. An alternative vector machine formulation was found to improve on the results produced by the established SVM in randomized bootstrap validation trials, yielding a receiver-operating characteristic curve area of 0.82 which represents a statistically significant improvement over the SVM technique in this application.
    Article · Jul 2013 · Journal of Digital Imaging
  • Source
    [Show abstract] [Hide abstract] ABSTRACT: It is recommended that BRCA1/2 mutation carriers undergo breast cancer screening using MRI because of their very high cancer risk and the high sensitivity of MRI in detecting invasive cancers. Clinical observations suggest important differences in the natural history between breast cancers due to mutations in BRCA1 and BRCA2, potentially requiring different screening guidelines. Three studies of mutation carriers using annual MRI and mammography were analyzed. Separate natural history models for BRCA1 and BRCA2 were calibrated to the results of these studies and used to predict the impact of various screening protocols on detection characteristics and mortality. BRCA1/2 mutation carriers (N = 1,275) participated in the studies and 124 cancers (99 invasive) were diagnosed. Cancers detected in BRCA2 mutation carriers were smaller [80% ductal carcinoma in situ (DCIS) or ≤10 mm vs. 49% for BRCA1, P < 0.001]. Below the age of 40, one (invasive) cancer of the 25 screen-detected cancers in BRCA1 mutation carriers was detected by mammography alone, compared with seven (three invasive) of 11 screen-detected cancers in BRCA2 (P < 0.0001). In the model, the preclinical period during which cancer is screen-detectable was 1 to 4 years for BRCA1 and 2 to 7 years for BRCA2. The model predicted breast cancer mortality reductions of 42% to 47% for mammography, 48% to 61% for MRI, and 50% to 62% for combined screening. Our studies suggest substantial mortality benefits in using MRI to screen BRCA1/2 mutation carriers aged 25 to 60 years but show important clinical differences in natural history. Impact: BRCA1 and BRCA2 mutation carriers may benefit from different screening protocols, for example, below the age of 40. Cancer Epidemiol Biomarkers Prev; 21(9); 1458-68. ©2012 AACR.
    Full-text available · Article · Jun 2012 · Cancer Epidemiology Biomarkers & Prevention
  • Article · Jun 2012 · Cancer Epidemiology Biomarkers & Prevention
  • Source
    K Passaperuma · E Warner · P A Causer · [...] · S A Narod
    [Show abstract] [Hide abstract] ABSTRACT: The addition of breast magnetic resonance imaging (MRI) to screening mammography for women with BRCA mutations significantly increases sensitivity, but there is little data on clinical outcomes. We report screening performance, cancer stage, distant recurrence rate, and breast cancer-specific mortality in our screening study. From 1997 to 2009, 496 women aged 25 to 65 years with a known BRCA1/2 mutation, of whom 380 had no previous cancer history, were enrolled in a prospective screening trial that included annual MRI and mammography. In 1847 screening rounds, 57 cancers were identified (53 screen-detected, 1 interval, and 3 incidental at prophylactic mastectomy), of which 37 (65%) were invasive. Sensitivity of MRI vs mammography was 86% vs 19% over the entire study period (P<0.0001), but was 74% vs 35% from 1997 to 2002 (P=0.02) and 94% vs 9% from 2003 to 2009 (P<0.0001), respectively. The relative sensitivities of MRI and mammography did not differ by mutation, age, or invasive vs non-invasive disease. Of the incident cancers, 97% were Stage 0 or 1. Of 28 previously unaffected women diagnosed with invasive cancer, 1 BRCA1 mutation carrier died following relapse of a 3 cm, node-positive breast cancer diagnosed on her first screen at age 48 (annual breast cancer mortality rate=0.5%). Three patients died of other causes. None of the 24 survivors has had a distant recurrence at a median follow-up of 8.4 years since diagnosis. Magnetic resonance imaging surveillance of women with BRCA1/2 mutations will detect the majority of breast cancers at a very early stage. The absence of distant recurrences of incident cancers to date is encouraging. However, longer follow-up is needed to confirm the safety of breast surveillance.
    Full-text available · Article · May 2012 · British Journal of Cancer
  • Peter Siegler · Claire M.B. Holloway · Petrina Causer · [...] · Donald B Plewes
    [Show abstract] [Hide abstract] ABSTRACT: To achieve high-quality unilateral supine breast magnetic resonance imaging (MRI) as a step to facilitate image aiding of clinical applications, which are often performed in the supine position. Contrast-enhanced breast MRI is a powerful tool for the diagnosis of cancer. However, prone patient positioning typically used for breast MRI hinders its use for image aiding. A fixture and a flexible four-element receive coil were designed for patient-specific shaping and placement of the coil in close conformity to the supine breast. A 3D spoiled gradient sequence was modified to incorporate compensation of respiratory motion. The entire setup was tested in volunteer experiments and in a pilot patient study. The flexible coil design and the motion compensation produced supine breast MR images of high diagnostic value. Variations in breast shape and in tissue morphology within the breast were observed between a supine and a diagnostic prone MRI of a patient. The presented supine breast MRI achieved an image quality comparable to diagnostic breast MRI. Since supine positioning is common in many clinical applications such as ultrasound-guided breast biopsy or breast-conserving surgery, the registration of the supine images will aid these applications.
    Article · Nov 2011 · Journal of Magnetic Resonance Imaging
  • [Show abstract] [Hide abstract] ABSTRACT: Objective: The addition of magnetic resonance imaging (MRI) to mammography for surveillance of women with BRCA mutations significantly increases sensitivity but lowers specificity. This study aimed to examine whether MRI surveillance, and particularly recall, is associated with increased anxiety, depression, or breast cancer worry/distress. Methods: Women with BRCA mutations in an MRI surveillance study were invited to complete: Hospital Anxiety and Depression Scale (HADS), Lerman's Breast Cancer Worry Scale, Breast Cancer Worry Interference Scale, and a quality of life rating at 3 time points: 1-2 weeks before (T1), 4-6 weeks after (T2) and 6 months after their annual surveillance (T3). Repeated measures analyses were performed over the 3 time points for recalled and non-recalled women. Results: 55 women (30 BRCA1, 25 BRCA2) completed study instruments at T1 and T2, and 48 at T3. Eighteen women (32%) were recalled for additional imaging. At T1, 27 women (49%) were above HADS threshold for "possible cases" for anxiety (score≥8). Recalled (but not non-recalled) women had a significant increase of HADS anxiety at T2 which dropped to below baseline by T3. No group differences were observed in terms of change over time in other quantitative psychological measures. Conclusions: While breast MRI surveillance did not have a detrimental psychological impact on women with a BRCA1 or BRCA2 mutation, recalling these very high-risk women for further imaging after a false positive MRI scan temporarily increased their global anxiety.
    Article · May 2011 · Breast (Edinburgh, Scotland)
  • K. Passaperuma · D. B. Plewes · P. Causer · [...] · E. Warner
    Article · May 2011 · Journal of Clinical Oncology
  • [Show abstract] [Hide abstract] ABSTRACT: To compare the underestimation of ductal carcinoma in situ (DCIS) vs DCIS with "possible invasion" at breast biopsy and to determine if any factors related to clinical indication, imaging abnormality, biopsy, or DCIS-grade affected the likelihood of underestimation. Of 3836 consecutive lesions that were biopsied by using a 14-gauge needle, 117 lesions revealed DCIS. Surgical pathology results of invasive carcinoma were compared with needle biopsy results of DCIS or DCIS with possible invasion. Clinical indication, imaging abnormality, biopsy guidance modality, sample number, and histologic grade were recorded. Yates corrected χ(2) and Fisher exact tests were used to determine differences between groups. A total of 101 lesions were DCIS and 16 were DCIS with possible invasion at biopsy. Thirty-six of 117 lesions (31%) revealed invasive carcinoma at resection pathology. Invasive carcinoma was present more often when DCIS with possible invasion was diagnosed compared with pure DCIS (7/16 [44%] vs 29/101 [29%], P = .36). No factor, including clinical indication, imaging abnormality, biopsy guidance method, sample number, or grade, was found to significantly affect the likelihood of underestimation for lesions diagnosed as DCIS vs DCIS with "possible invasion." The likelihood of pure DCIS underestimation significantly increased when lesions were high grade compared with either intermediate or low grade (18/44 [41%] vs 9/44 [21%] vs 2/10 [20%], P = .03). For lesions biopsied by using a 14-gauge needle, there is a trend towards underestimation of the presence of invasive carcinoma when pathology reveals DCIS with possible invasion compared with pure DCIS. High-grade DCIS was significantly more likely to be underestimated.
    Article · May 2011 · Canadian Association of Radiologists Journal
  • Ellen Warner · Kimberley Hill · Petrina Causer · [...] · Steven A Narod
    [Show abstract] [Hide abstract] ABSTRACT: The sensitivity of magnetic resonance imaging (MRI) for breast cancer screening exceeds that of mammography. If MRI screening reduces mortality in women with a BRCA1 or BRCA2 mutation, it is expected that the incidence of advanced-stage breast cancers should be reduced in women undergoing MRI screening compared with those undergoing conventional screening. We followed 1,275 women with a BRCA1 or BRCA2 mutation for a mean of 3.2 years. In total, 445 women were enrolled in an MRI screening trial in Toronto, Ontario, Canada, and 830 were in the comparison group. The cumulative incidences of ductal carcinoma in situ (DCIS), early-stage, and late-stage breast cancers were estimated at 6 years in the cohorts. There were 41 cases of breast cancer in the MRI-screened cohort (9.2%) and 76 cases in the comparison group (9.2%). The cumulative incidence of DCIS or stage I breast cancer at 6 years was 13.8% (95% CI, 9.1% to 18.5%) in the MRI-screened cohort and 7.2% (95% CI, 4.5% to 9.9%) in the comparison group (P = .01). The cumulative incidence of stages II to IV breast cancers was 1.9% (95% CI, 0.2% to 3.7%) in the MRI-screened cohort and 6.6% (95% CI, 3.8% to 9.3%) in the comparison group (P = .02). The adjusted hazard ratio for the development of stages II to IV breast cancer associated with MRI screening was 0.30 (95% CI, 0.12 to 0.72; P = .008). Annual surveillance with MRI is associated with a significant reduction in the incidence of advanced-stage breast cancer in BRCA1 and BRCA2 carriers.
    Article · Mar 2011 · Journal of Clinical Oncology
  • Petrina A Causer
    [Show abstract] [Hide abstract] ABSTRACT: Breast magnetic resonance imaging (MRI) is indisputably the highest sensitivity test available to detect breast cancer, revealing more extensive cancer in the ipsilateral and otherwise occult cancer in the contralateral breasts when used before surgery. The use of preoperative breast MRI has become somewhat controversial, because the clinical benefit of the heightened detection provided by MRI has been questioned in the context of multidisciplinary breast cancer treatment, relatively low local recurrence, and metachronous contralateral cancer rates. Also, MRI detection rates have been compared with the high rates reported in the pathology literature. The emerging clinical outcome literature is showing conflicting results to demonstrating actual overall benefit. Critical review of this literature reveals several misconceptions about MRI detection rates and limitations of many of the published outcome studies to date, which render the results not necessarily generalizable to contemporary optimized breast MRI practices. This article addresses some of the misconceptions raised by critics, provides a critical review of the clinical outcome literature, reviews patient subgroups anticipated to have the highest yield when using preoperative MRI, makes recommendations for optimizing breast MRI practice, and suggests areas for potential future research.
    Article · Feb 2011 · Canadian Association of Radiologists Journal
  • [Show abstract] [Hide abstract] ABSTRACT: Breast MRI is often used for surveillance of breast cancer (BC) survivors despite the lack of evidence in this population. We surveyed younger BC survivors to evaluate their willingness to participate in a randomized controlled trial (RCT) of annual digital mammography with or without MRI. Median age of the 348 participants was 51 years; 45% had undergone diagnostic MRI. 22% continued to have surveillance MRI. 58% agreed to consider participating in the proposed RCT; 16% remained neutral. An RCT of MRI surveillance for BC survivors <age 60 is still feasible at centres where annual MRI is not yet routinely recommended.
    Article · Feb 2011 · Breast (Edinburgh, Scotland)
  • Ellen Warner · Petrina A Causer · John W-N Wong · [...] · Donald B Plewes
    [Show abstract] [Hide abstract] ABSTRACT: Although magnetic resonance imaging (MRI) is much more sensitive than mammography for detecting early invasive breast cancer, in many high-risk screening studies MRI was less sensitive than mammography for detecting ductal carcinoma in situ (DCIS). We reviewed our experience detecting DCIS in our single center study of annual MRI, mammography, ultrasound and clinical breast examination (CBE) for screening very high-risk women. All cases of DCIS±microinvasion and invasive cancer were compared in two time frames: before (period A) and after (period B) July 2001-when we acquired expertise in the detection of DCIS with MRI-with respect to patient demographics, method of detection, and rates of detection of invasive cancer and DCIS. In period A there were 15 cases (3.1% of 486 screens) in 223 women, of which 2 (13%) were DCIS-one with microinvasion-neither detected by MRI. In period B there were 29 cases (3.3% of 877 screens) in 391 women, of which 10 (34%) were DCIS±microinvasion (p=0.04), all 10 detected by MRI but only one by mammography. No DCIS cases were detected by ultrasound or CBE. Specificity was lower in period B than in period A but acceptable. The ability to detect DCIS with screening MRI improves significantly with experience. MRI-guided biopsy capability is essential for a high-risk screening program. In experienced centers the increased sensitivity of MRI relative to mammography is at least as high for DCIS as it is for invasive breast cancer.
    Article · Jan 2011 · The Breast Journal
  • Maude Labelle · Petrina A. Causer · Ravi Menezes · [...] · Andrew Nelson
    [Show abstract] [Hide abstract] ABSTRACT: PURPOSE To determine the frequency of missing the target lesion at MRI-guided vacuum-assisted biopsy (VAB) at our institution, how often missed lesions were identified at the time of biopsy and follow-up imaging and to evaluate whether any lesion or breast characteristics affect missing the target. METHOD AND MATERIALS With institutional Research Ethics Board approval, using a database, 277 consecutive lesions in 249 patients that underwent MRI-guided VAB, with final surgical pathology or at least six months MRI follow-up for concordant benign pathology were examined. Biopsies were performed in a 1.5T magnet using a 9 gauge MRI-compatible VAB device and localization software. Frequency of missed lesions recognized prospectively at the time of biopsy or follow-up MRI and on retrospective review of the MR images was determined. The association between a lesion being missed on biopsy and lesion or breast characteristics was examined using Chi-square and Mann-Whitney tests. All tests were two-sided using 0.05 for significance. RESULTS Among 277 biopsied lesions, 18 (6%) were missed based on retrospective review of the MR images from the time of biopsy or follow-up. 17 missed lesions were followed on MRI; one missed lesion was surgically excised due to discordance. Of the missed lesions, 2/18 (11%) were identified prospectively at time of biopsy and 11/17 (65%) prospectively on MRI follow-up. On retrospective imaging review 4/18 (22%) were found to be missed at the time of biopsy and 14/17 (82%) on follow-up imaging. 5/18 (28%) missed lesions had surgical excision pathology including 1/18 (6%) invasive ductal carcinoma, detected at six month follow-up and 1/18 (6%) lobular carcinoma in situ, the remaining benign based on follow-up MRI or surgical pathology. No lesion or breast characteristics significantly affected the likelihood of missing a lesion. CONCLUSION The target lesion was missed in 6% of MRI-guided VAB, of these only 11% were recognized at the time of biopsy. No factors signicantly affected biopsy sampling accuracy. Follow-up MRI is required following a concordant benign pathology result to determine if the lesion was sampled. An initial follow-up interval of six months can detect change indicative of malignancy. CLINICAL RELEVANCE/APPLICATION Due to inaccuracy assessing lesion sampling at the time of MRI-guided VAB an initial 6 month follow-up MRI is mandatory for concordant benign pathology results.
    Conference Paper · Nov 2010
  • Maude Labelle · Petrina A. Causer · Roberta A. Jong · [...] · Andrew Nelson
    [Show abstract] [Hide abstract] ABSTRACT: PURPOSE/AIM The purpose of this exhibit is to review the indications and technique of performing MRI-guided vacuum-assisted breast biopsy and identify potential pitfalls that can lead to: 1) Inability to perform the procedure 2) Missing the target lesion 3) Inability to accurately assess lesion sampling based on MR images CONTENT ORGANIZATION 1. Review MRI guided biopsy indications 2. Summarize MRI-guided biopsy technique and complications 3. Pre-procedural associated pitfalls: Biopsy planning Imaging interpretation: diagnostic vs biopsy examinations 4. Procedural associated pitfalls: Computer assisted targeting Imaging interpretation: Accuracy of needle placement and post-biopsy/clip imaging. 5. Post-procedural associated pitfalls: Radiologic-pathologic correlation MR imaging follow-up requirement SUMMARY According to the American College of Radiology Breast MRI Accreditation Program, MRI-guided vacuum-assisted breast biopsy is a mandatory component of any breast MRI practice. Despite the maturation of MR-guided biopsy there are many pitfalls that can compromise biopsy success. Recognizing the potential pitfalls and how to avoid them can ensure adequate lesion sampling and optimal patient care.
    Conference Paper · Nov 2010
  • Source
    F.C. Wright · J Zubovits · S Gardner · [...] · P Causer
    [Show abstract] [Hide abstract] ABSTRACT: Accurate assessment of residual disease after neo-adjuvant chemotherapy (NEC) for women with locally advanced and inflammatory breast cancer (LABC) is critical for planning surgery. The study's purpose was to prospectively determine the optimal method (clinical examination (CE), mammogram (MG), and magnetic resonance imaging (MRI)) for assessing residual disease after NEC for women with LABC. Women with LABC who received NEC and surgery were enrolled. Patient demographics, tumor size as measured by CE, MG, and MRI both before and after NEC, and final pathologic size of tumor were collected. Response to NEC was calculated using RECIST criteria. Paired t-tests and the Pearson correlation were used to compare tumor size on CE, MG, MRI, and final pathology. Forty-eight women with 50 LABC were recruited. Mean pre-NEC tumor size was 8.2, 5.1, and 6.2 cm on CE, MG, and MRI. Mean post-NEC tumor size was 2.4, 4.3, 3.9, and 3.6 cm on CE, MG, MRI, and final pathology. The Pearson correlation co-efficient between post-NEC measurements and pathology was 0.63 (CE), 0.15 (MG), and 0.49 (MRI). We found that there was limited correlation between the extent of residual disease after NEC for patients with LABC as assessed by CE, MG, and MRI as compared to final pathology.
    Full-text available · Article · Jun 2010 · Journal of Surgical Oncology
  • E. Warner · K. Hill · P. Causer · [...] · S. Narod
    Article · Dec 2009 · Cancer Research
  • T Arazi-Kleinman · P.A. Causer · R.A. Jong · [...] · E Warner
    [Show abstract] [Hide abstract] ABSTRACT: To evaluate the sensitivity and specificity of magnetic resonance imaging (MRI) computer-aided detection (CAD) for breast MRI screen-detected lesions recommended for biopsy in a high-risk population. Fifty-six consecutive Breast Imaging Reporting and Data System (BI-RADS) 3-5 lesions with histopathological correlation [nine invasive cancers, 13 ductal carcinoma in situ (DCIS) and 34 benign] were retrospectively evaluated using a breast MRI CAD prototype (CAD-Gaea). CAD evaluation was performed separately and in consensus by two radiologists specializing in breast imaging, blinded to the histopathology. Thresholds of 50, 80, and 100% and delayed enhancement were independently assessed with CAD. Lesions were rated as malignant or benign according to threshold and delayed enhancement only and in combination. Sensitivities, specificities, and negative predictive values (NPV) were determined for CAD assessments versus pathology. Initial MRI BI-RADS interpretation without CAD versus CAD assessments were compared using paired binary diagnostic tests. Threshold levels for lesion enhancement were: 50% to include all malignant (and all benign) lesions; and 100% for all invasive cancer and high-grade DCIS. Combined use of threshold and enhancement patterns for CAD assessment was best (73% sensitivity, 56% specificity and 76% NPV for all cancer). Sensitivities and NPV were better for invasive cancer (100%/100%) than for all malignancies (54%/76%). Radiologists' MRI interpretation was more sensitive than CAD (p=0.05), but less specific (p=0.001) for cancer detection. The breast MRI CAD system used could not improve the radiologists' accuracy for distinguishing all malignant from benign lesions, due to the poor sensitivity for DCIS detection.
    Article · Dec 2009 · Clinical Radiology
  • Source
    Jacob E.D. Levman · Petrina Causer · Ellen Warner · Anne L Martel
    [Show abstract] [Hide abstract] ABSTRACT: To evaluate the effect that variations in the enhancement threshold have on the diagnostic accuracy of two computer-aided detection (CAD) systems for magnetic resonance based breast cancer screening. Informed consent was obtained from all patients participating in cancer screening and this study was approved by the participating institution's review board. This retrospective study was nested in a prospective, single-institution, high-risk, breast screening study involving dynamic contrast-enhanced magnetic resonance imaging. Only those screening examinations (n = 223) for which a histopathological diagnosis was available were included. Two CAD methods were performed: the signal enhancement ratio (SER) and support vector machines (SVMs). Statistical analysis was performed by tracking changes in each CAD test's diagnostic accuracy (eg, receiver-operating characteristic [ROC] curve area, maximum possible sensitivity) with changes in the enhancement threshold. The enhancement threshold plays a significant role in affecting a CAD test's potential sensitivity, ROC curve area, and number of assumed true and false-positive predictions per cancerous examination. A high threshold can also limit the CAD-based detection of the full size of a lesion. Enhancement thresholds can limit a CAD test's ability to diagnose a lesion's full size and as such should not be raised above 60%. The clinically used SER method exhibits a high rate of false positives at low enhancement thresholds and as such the threshold should not be set lower than 50%. The SVM method yielded better results in our study than the SER method at clinically realistic enhancement thresholds.
    Full-text available · Article · Jul 2009 · Academic radiology