Sandra Catarzi

Comparison of standard reading and computer aided detection (CAD) on a national proficiency test of screening mammography

OBJECTIVE To evaluate the role of computer aided detection (CAD) in improving the interpretation of screening mammograms MATERIAL AND METHODS Ten radiologists underwent a proficiency test of screening mammography first by conventional reading and then with the help of CAD. Radiologists were blinded to test results for the whole study duration. Results of conventional and CAD reading were compared in terms of sensitivity and recall rate. Double reading was simulated combining conventional readings of four expert radiologists and compared with CAD reading. RESULTS Considering all ten readings, cancer was identified in 146 or 153 of 170 cases (85.8 vs. 90.0%; chi(2)=0.99, df=1, P=0.31) and recalls were 106 or 152 of 1330 cases (7.9 vs. 11.4%; chi(2)=8.69, df=1, P=0.003) at conventional or CAD reading, respectively. CAD reading was essentially the same (sensitivity 97.0 vs. 96.0%; chi(2)=7.1, df=1, P=0.93; recall rate 10.7 vs. 10.6%; chi(2)=1.5, df=1, P=0.96) as compared with simulated conventional double reading. CONCLUSION CAD reading seems to improve the sensitivity of conventional reading while reducing specificity, both effects being of limited size. CAD reading had almost the same performance of simulated conventional double reading, suggesting a possible use of CAD which needs to be confirmed by further studies inclusive of cost-effective analysis.

Independent Double Reading of Screening Mammograms

Objective — To evaluate the cost effectiveness of independent double reading of screening mammograms. Setting — Prospective study of 18817 women undergoing first or repeat screening in a population based programme in the Florence district. Methods — Mammograms were independently double read by experienced radiologists. Subjects with mammographic abnormalities reported by at least one reader were recalled for diagnostic assessment. The mean increase in recall rate, cancer detection rate, and screening costs attributable to double reading was calculated. Results — Eleven of 125 cancers were detected by only one reader. The mean increase in cancer detection rate attributable to double reading compared with single reading was 4·6% (95% confidence interval (CI) 1·1 to 8·9). From a total of 748 cases referred for diagnostic assessment, 196 subjects were referred by one reader only. The mean increase in referral rate attributable to double reading compared with single reading was 15·1% (CI 12·3 to 17·8). Double reading caused a marked increase in the cost for each woman screened −8·5% at the first screening and 6·2% at repeat screening and a more limited increase in the cost for each cancer detected −3·5% at the first screening and 2·7% at repeat screening. Cancers detected by only one screener were at an earlier stage than those detected by both screeners (P = 0·6, not significant). Conclusions — Independent double reading results in only a modest increase in the detection of cancers and therefore may not be cost effective.

Solid nonpalpable breast lesions Success and failure of guided fine-needle aspiration cytology in a consecutive series of 2444 cases

Purpose: To evaluate the contribution of guided fine-needle aspiration cytology in reducing unnecessary biopsies of benign solid nonpalpable breast lesions with low suspicion of malignancy at mammography. Material and Methods: An evaluation was made of a consecutive series of 2444 solid nonpalpable breast lesions detected by mammography and undergoing guided (sonography or stereotaxy) fine-needle aspiration cytology. Surgical biopsy was made in the presence of strong suspicion of malignancy at mammography and/or of abnormal cytology. Results: The sensitivity was 96.7% and the specificity 77.7% (average follow-up 2.77 years). False-negative/inadequate cytology associated with low suspicion of malignancy at mammography resulted in a diagnostic delay in 27 cancer cases (invasive 20, intraductal 7). On the other hand, cytology led to surgical biopsy in 53 cancer cases which might not otherwise have been biopsied because of low radiological suspicion of cancer. Surgical biopsy of all cases, to avoid diagnostic delays, would have increased the benign biopsy rate by a factor of 4.5, with a rise in the benign: malignant biopsy ratio from 0.44:1 to 1.93:1. Conclusion: Stereotaxy- or ultrasound-guided fine-needle aspiration cytology of nonpalpable mammographic abnormalities can achieve a sharp reduction in unnecessary benign biopsies in cases of low suspicion of malignancy at mammography.

NEGLIGIBLE ADVANTAGES AND EXCESS COSTS OF ROUTINE ADDITION OF BREAST ULTRASONOGRAPHY TO MAMMOGRAPHY IN DENSE BREASTS

Aim To assess the role of breast ultrasonography as a complement to negative mammography in radiologically dense breasts. Material and methods Out of a total series of 49,044 consecutive mammograms reported as negative in asymptomatic women, 25,665 (52.3%) were coded as dense (BI-RADS D3–4) and ultrasonography was recommended. Due to organizational problems, ultrasonography was performed immediately or within 1 month only in 5,227 cases, representing the study series. Results Two cancers were detected at immediate ultrasonography (0.03%). The cancer detection rate in women aged 40–49 and 50–69 years was 0.002% and 0.07%, respectively. The benign biopsy rate was 0.5% for core biopsies and 0.02% for surgical biopsies. The cost per ultrasonography-assessed woman was € 56.05, whereas the cost per additional mammographically occult but ultrasonography-detected cancer was € 146,496.53. The mammograms of the 2 cancer cases underwent blind review by an expert reader and were confirmed as negative. Discussion Our findings show a low cancer detection rate, substantially lower compared to other clinical studies of ultrasonography in dense breasts, though in accordance with preliminary evidence from an Italian randomized clinical trial within a population-based screening program. The policy of adding ultrasonography to negative mammography in dense breasts seems to have very limited cost-effectiveness, and should not be adopted in routine practice before results of ongoing clinical trials are available.

Breast fine-needle aspiration malondialdehyde deoxyguanosine adduct in breast cancer

Abstract This study has analysed the generation of 3-(2-deoxy-β-D-erythro-pentafuranosyl)pyrimido[1,2-α]purin-10(3H)-one deoxyguanosine adduct [M1dG], a biomarker of oxidative stress and lipid peroxidation, in breast fine-needle aspirate samples of 22 patients with breast cancer, at different clinical stages, in respect to 13 controls. The multivariate analysis show that M1dG adduct was higher in cases than in controls (Mean Ratio (MR) = 5.26, 95% CI = 3.16–8.77). Increased M1dG was observed in women with a tumour grade 3 and a pathological diameter 2 (MR = 7.61, 95% CI = 3.91–14.80 and MR = 5.75, 95% CI = 3.13–10.59, respectively). A trend with increasing tumour grade and pathological diameter was present (MR = 1.98, 95% CI = 1.57–2.50 and MR = 2.44, 95% CI = 1.71–3.48, respectively). Not significant effects of age and smoking habit were found (MR = 1.58, 95% CI = 0.92–2.72 and MR = 1.68, 95% CI 0.88–3.20, respectively). An increment over the background frequency of M1dG can contribute to breast cancer development. Increasing severity of breast tumour can influence DNA damage level.

The role of arbitration of discordant reports at double reading of screening mammograms

Objective: To asses the effectiveness of arbitration of discordant double readings in mammography screening. Design: A retrospective study of 1217 consecutive arbitrations. Setting: A subset of discordant double readings from the Florence screening programme underwent arbitration by a third reader. Results: Positive arbitration of 1217 discordant double readings prompted assessment in 476 cases (39.2%), detecting 30 cancers (6.3%). Of 741 negative arbitrations (60.8%), 311 have been followed up thus far, and two cancers (0.64%) occurred in the site previously suspected at one of the two independent readings. Arbitration had a sensitivity of 86.3% and a negative predictive value of 99.3%. Arbitration reduced the overall referral rates from 3.82% to 2.59% (relative decrease 32.1%). Due to false-negative arbitration, cancers detected per 1000 women screened would decrease from 4.58 to 4.50 (relative decrease 1.7%). For every cancer missed due to false-negative arbitration, 151 unnecessary recalls and €21,248 would have been saved, whereas the saved cost per screened woman due to arbitration was €1.72. Discussion: Arbitration of discordant double reading would substantially reduce referral rates with a limited reduction in cancer detection rate, and may be recommended as a routine procedure. Greater benefit from arbitration might be expected in the presence of high referral rates at independent double reading, a common scenario in a newly implemented service screening.

Accuracy of needle biopsy of breast lesions visible on ultrasound: audit of fine needle versus core needle biopsy in 3233 consecutive samplings with ascertained outcomes.

INTRODUCTION Core needle biopsy (CNB) has progressively replaced fine needle aspiration cytology (FNAC) in the diagnosis of breast lesions. Less information is available on how these tests perform for biopsy of ultrasound (US) visible breast lesions. This study examines the outcomes of CNB and FNAC in a large series ascertained with surgical histology or clinical-imaging follow-up. MATERIALS AND METHODS Retrospective five-year audit of 3233 consecutive US-guided needle samplings of solid breast lesions, from self-referred symptomatic or asymptomatic subjects, performed by six radiologists in the same time-frame (2003-2006): 1950 FNAC and 1283 CNB. The probability of undergoing CNB as a first test instead of FNAC was evaluated using logistic regression. Accuracy and inadequacy were calculated for each of CNB and FNAC performed as first test. Accuracy measures included equivocal or borderline/atypical lesions as positive results. RESULTS The probability of CNB as a first test instead of FNAC increased significantly over time, when there was a pre-test higher level of suspicion, in younger (relative to older) women, with increasing lesion size on imaging, and for palpable (relative to impalpable) lesions. Inadequacy rate was lower for CNB (B1 = 6.9%) than for FNAC (C1 = 17.7%), p < 0.001, and specifically in malignant lesions (B1 = 0.9% vs. C1 = 4.5%; p < 0.001). False negative rate was equally low for both CNB and FNAC (1.7% each test). CNB performed significantly better than FNAC for absolute sensitivity (93.1% vs. 74.4%; p < 0.001) and complete sensitivity (97.4% vs. 93.8%; p = 0.001), however specificity was lower for CNB than FNAC (88.3% vs. 96.4%; p < 0.001). Absolute diagnostic accuracy was higher for CNB than FNAC (84.5% vs. 71.9; p < 0.001) while FNAC performed better than CNB for complete diagnostic accuracy (95.4% vs. 93.2; p < 0.008). In the small subgroup assessed with CNB after an inconclusive initial FNAC (231 cases) there was improved complete sensitivity (from 93.8% to 97.0%) however this also increased costs. CONCLUSION FNAC and CNB were generally performed in different patients, thus our study reported indirect comparisons of these tests. Although FNAC performed well (except for relatively high inadequacy), CNB had significantly better performance based on measures of sensitivity, but this was associated with lower specificity for CNB relative to FNAC. Overall, CNB is the more reliable biopsy method for sonographically-visible lesions; where FNAC is used as the first-line test, inadequate or inconclusive FNAC can be largely resolved by using repeat sampling with CNB.

The role of galactography in the detection of breast cancer.

The authors report on a series of 529 consecutive patients examined on physical examination, mammography, nipple discharge cytology and galactography. The criterion for galactography was essentially bloody nipple discharge (73% of cases). Serous nipple discharge was not considered worthy of routine galactography since it is associated with an extremely low incidence of breast cancer. Surgical excision and histologic examination of the discharging duct was performed in 200 cases. Eighteen cases of breast cancer were detected (10 infiltrating, 8 intraductal) of which 9, 6, 7 or 7 were suspected on physical examination, mammography, cytology or galactography, respectively. All combined tests suspected 13 of 18 breast cancers; 3 intraductal breast cancers were biopsied because of evidence of multiple papillomas on galactography, and 2 infiltrating breast cancers were operated because of persistent bloody nipple discharge in the absence of any other sign. No breast cancer was suspected on galactography alone. Galactography is indicated in the presence of bloody nipple discharge, and a biopsy should be performed when breast cancer or multiple papillomas are suspected. The diagnosis and excision of a single papilloma (breast cancer was never misdiagnosed as a single papilloma on galactography) is not worthwhile since a single papilloma is a benign lesion, and the benefit of its excision is still unclear.

Florence–Sydney Breast Biopsy Study: sensitivity of ultrasound-guided versus freehand fine needle biopsy of palpable breast cancer

Background. Fine needle aspiration biopsy (FNAB) is widely used in the diagnosis of breast cancer. It is unknown whether, for palpable cancers, ultrasound-guided FNAB is more accurate than freehand FNAB, and practice varies between physicians, services and countries.Methods. From consecutive women attending a major cancer centre in Florence, we prospectively recruited subjects who had a definitely palpable lump which was solid on ultrasound and suspicious of malignancy (n=102). All subjects were investigated using both ultrasound-guided and freehand FNAB (one aspirate with each method). Radiologists skilled in both sampling techniques performed all clinical examinations and aspirations, and for each subject the same radiologist obtained both FNAB samples. Sequence of aspiration method was randomised. Cytological interpretation was blinded to method of sampling. Comparative sensitivity (and insufficiency) for FNAB using the two methods was calculated in all cancers (n=97).Results. Ultrasound-guided FNAB resulted in 13.6% (5–22%) less insufficient aspirates than freehand FNAB (χ2=7.58; p=0.006). When insufficient aspirates are included and considered as negative, ultrasound-guided FNAB has a 14.6% (5.8–23%) or a 16.5% (7.6–25.4%) significantly better sensitivity than freehand FNAB (for cytology 3–5 positive or cytology 4–5 positive respectively). When insufficient aspirates are excluded from the analysis, ultrasound-guided FNAB has a 1.4% (−1.2 to 3.9%) or a 2.6% (−2.5 to 7.8%) higher sensitivity than freehand FNAB (for cytology 3–5 positive or cytology 4–5 positive respectively) but this difference in the sensitivity of the two methods is not statistically significant.Conclusion. Our data suggest that ultrasound-guided FNAB has better sensitivity than freehand FNAB in palpable breast cancer, which is predominantly an effect of a significant reduction in insufficient aspirates, but in part an effect of ‘upgrading’ cytological classification of cancers.

Proficiency test for screening mammography: results for 117 volunteer Italian radiologists

Objective To analyse the performance of a large sample of Italian radiologists undergoing a proficiency test for screening mammography. Design Evaluation of performance indicators according to reference standards determined by a panel of experts (sensitivity (reference standard ≥80%), recall rate (reference standard ≤15%)). Setting 117 Italian radiologists of varying experience (years of practice 0.5–18, average 5.9; mammograms read 500–51 000, average 13 000), all currently reporting clinical mammography and planning to take part in screening in the near future. Results Eighty four of 117 (72%) radiologists reached the standard for sensitivity, 88 (75%) reached the standard for recall rate, and only 59 (50%) reached both standards and passed the proficiency test. The probability of passing the test was significantly correlated with mammographic practice (p=0.015), mammograms read (p=0.024), and mammograms read/year (p=0.043). Discussion The performance of a large sample of Italian radiologists currently reporting clinical mammography was disappointing, indicating the need for proper training of at least 50% of the tested subjects. When implementing organised screening the health authority should set up a proper process for training and accrediting radiologists, and a proficiency test should be part of such a process.