J.H.C.L. Hendriks

Histologic multifocality of Tis, T1-2 breast carcinomas: implications for clinical trials of breast-conserving surgery

Breast cancer multifocality was studied in mastectomy specimens by correlated specimen radiography and histologic techniques. The patients chosen for study were comparable to those eligible for breast‐conserving surgical therapy. Two study groups, one with 282 invasive cancers (T1‐2) and the other with 32 intraductal cancers, were selected from a group of 399 consecutive cases by omitting patients who were clearly, or very probably, not candidates for breast‐conserving surgical therapy according to current trial criteria. Omitted patients included those with clinically and/or radiologically multifocal cancers and patients with tumor extension into the chest wall or skin (7%). Also excluded were the socalled diffuse invasive cancers (8%), the clinically and radiologically occult tumors (3%), and the invasive cancers larger than 5 cm (3%). Of the 282 invasive cancers, 105 (37%) showed no tumor foci in the mastectomy specimen around the reference mass. In 56 (20%) tumor foci were present within 2 cm, and in 121 (43%) tumor was found more than 2 cm from the reference tumor. In 75 (27%) the tumor foci beyond 2 cm were histologically noninvasive cancers, and in 46 cases (16%) they contained invasive cancers as well. A comparison between the group with reference tumors less than 2 cm and the group with reference tumors more than 2 cm in size showed no significant difference between the groups in terms of presence or absence of tumor foci or distance of tumor foci from the reference tumor. If the 264 invasive cancers in this series that were 4 cm or less in diameter had been removed with a margin of 3 to 4 cm, 7% to 9% of the patients would have had invasive cancer left in the remaining breast tissue, and 4% to 9% would have had foci of noninvasive cancer left in the remaining breast tissue. On the basis of the data on the distribution of tumor at different distances from the reference tumor, the current study estimates the expected rates of local recurrences after breast‐conserving surgical procedures relative to the extensiveness of the excision. The possible impact of postoperative local radiation therapy on the rates of expected local recurrence is discussed.

Extent, distribution, and mammographic/ histological correlations of breast ductal carcinoma in situ

To assess the potential of breast-conserving treatment for ductal carcinoma in situ (DCIS), 82 mastectomy specimens were studied by Egans serial subgross method. 42 (51%) of the tumours were larger than 50 mm and only 12 (15%) were smaller than 20 mm; the size distribution was not affected by the mode of detection (mammography 52 cases, clinical examination 30). All but 1 case showed only 1 region of tumour. 66% of tumours involved one breast quadrant, 23% extended over more than one quadrant, and 11% were centrally located. Mammographic estimates, based on the extent of microcalcifications, frequently underestimated the histological size of tumours, the extent of the discrepancy being related to the histological type--8/50 predominantly micropapillary/cribriform. In view of the frequently large size, adequate excision of many DCIS will require a wide excision involving up to a whole quadrant.

REDUCTION OF BREAST CANCER MORTALITY THROUGH MASS SCREENING WITH MODERN MAMMOGRAPHY: First Results of the Nijmegen Project, 1975 - 1981

Since 1975 four rounds of screening with modern mammography for breast cancer have been carried out among 30 000 Nijmegen women born before 1940. The results up to the end of 1981 shows that the odds ratio of screened vs unscreened subjects among women who died from breast cancer compared with women who did not, was 0.48 (95% confidence interval 0.23-1.00) in all age groups.

Initiation of population-based mammography screening in Dutch municipalities and effect on breast-cancer mortality: a systematic review.

BACKGROUND More than a decade ago, a mammography screening programme for women aged 50-69 years was initiated in the Netherlands. Our aim was to assess the effect of this programme on breast-cancer mortality rates. METHODS We examined data for 27948 women who died of breast-cancer aged 55-74 years between 1980 and 1999 (30560 cases until 2001). We grouped individuals into 93 clusters, depending on where they lived, and analysed data by use of national population statistics. We analysed time trends in breast-cancer mortality, adjusting for gradual implementations at municipality level, taking as year 0 the month and year in which screening began in a particular municipality. We used a Poisson regression model to estimate the time at which the trend started to turn. We assessed indirectly whether this turning point was related to initiation of screening or adjuvant systemic therapy in four clusters defined according to when screening was implemented. FINDINGS Compared with rates in 1986-88, breast-cancer mortality rates in women aged 55-74 years fell significantly in 1997 and subsequent years as predicted, reaching -19.9% in 2001. Mortality rates had been increasing by an annual 0.3% until screening was introduced; thereafter we noted a decline of 1.7% per year (95% CI 2.39-0.96) in women aged 55-74 years and of 1.2% in those aged 45-54 (2.40 to 0.07). The turning point in mortality trends arose at around year 0. Adjuvant systemic therapy is unlikely to be the cause of this turning point, since the mortality rates continued to rise up to 1 year after implementation in municipalities where screening began after 1995. INTERPRETATION Routine mammography screening can reduce breast-cancer mortality rates in women aged 55-74 years.

Age-dependent growth rate of primary breast cancer

Background. Currently, there have been no reports of clear‐cut beneficial effects of regular screening with mammography at a mean screening interval of 2 years in women younger than 50 years of age. It may be that, because of a higher growth rate of breast cancer, more frequent screening is necessary if any effect is to be observed in this age group. However, good quantitative estimates for the growth rate in different age groups are lacking.

The current detectability of breast cancer in a mammographic screening program. A review of the previous mammograms of interval and screen-detected cancers

Background. The occurrence of an interval cancer in a mammographic screening program is indicative of a suboptimum effect on mortality, because the very aim of the screening is to detect as many cancers as possible and at their earliest possible stage. In several studies, the previous screening mammograms of patients with an interval cancer were reviewed and the reasons for the “missed diagnosis” were classified into four categories: “screening error” (20–29%), “minimal sign present” (30–40%), “radiographically occult” (33–58%), or “radiographically occult at diagnosis” (occult both at previous screening and diagnosis; 7–16%). A similar procedure was followed in the Nijmegen screening project with patients recently diagnosed as having interval cancer or screen‐detected cancer.

Nation wide breast cancer screening in the Netherlands: Results of initial and subsequent screening 1990-1995

Based on an extensive cost‐effectiveness analysis, the Dutch nation‐wide breast cancer screening programme started in 1990, providing a biennial screen examination to women aged 50 to 69 years. The programme is monitored by the National Evaluation Team, which annually collects tabulated regional evaluation data to determine performance indicators. This study presents (trends in) the outcomes of initial and subsequent screening rounds, 1990–1995, and compares them to the predictions of the cost‐effectiveness‐analysis. Up to 1996, 88% of the target population was covered by the programme and more than 2.4 × 106 women were invited. The overall attendance rate was 77.5% with little differences between screening rounds and age groups; the highest rate was found in non‐urbanised areas (82.4%). Of 1,000 initially (and 2 years thereafter) screened women, 13.4 (6.6) were referred for further investigation, 9.7 (4.4) were biopsied and 6.4 (3.4) had breast cancer. The positive predictive values of screen test and biopsy were 47% (51%) and 66% (78%), respectively. DCIS was diagnosed in 0.9 (0.5) and invasive cancers ≤10 mm in 1.5 (1.0) per 1,000 screens. Lymph node metastases were found in 28% (24%) of the invasive cancers. Except the increasing attendance, which was much higher than expected, the results were fairly constant over the years. Contrary to initial screens, the results of subsequent screens did not fulfil expectations with regard to breast cancer detection and tumour size distribution. We conclude that the nation‐wide screening programme is being implemented successfully. Given the results, the programme should contribute to a substantial breast cancer mortality reduction in the future. The discrepancy between observed and expected results in subsequent screens has to be watched carefully. Int. J. Cancer 75:694–698, 1998.© 1998 Wiley‐Liss, Inc.

Mammographically occult breast cancer: A pathologic and radiologic study

Fifteen mammographically occult breast cancers, 3 of which were preinvasive and 12 invasive, and 52 breast cancers, which presented only microcalcifications mammographically without an associated tumor shadow (33 preinvasive and 19 invasive), were reviewed pathologically and radiologically. This study suggests that most of the preinvasive cancers, without mammographically significant calcifications, are, and some of the invasive cancers situated in dense breasts may be, mammographically occult. The mean diameter of the five mammographically occult invasive ductal carcinomas and the five invasive lobular carcinomas was 20 mm and 50 mm, respectively. Histologic characteristics, which may explain why invasive lobular carcinomas could be occult mammographically, even in an advanced stage, include a diffuse invasive pattern and, frequently, a poor desmoplastic reaction in contrast to the invasive ductal carcinomas. The implications of mammographically occult cancers on management of symptomatic and asymptomatic patients are discussed.

An automatic method to discriminate malignant masses from normal tissue in digital mammograms.

Specificity levels of automatic mass detection methods in mammography are generally rather low, because suspicious looking normal tissue is often hard to discriminate from real malignant masses. In this work a number of features were defined that are related to image characteristics that radiologists use to discriminate real lesions from normal tissue. An artificial neural network was used to map the computed features to a measure of suspiciousness for each region that was found suspicious by a mass detection method. Two data sets were used to test the method. The first set of 72 malignant cases (132 films) was a consecutive series taken from the Nijmegen screening programme, 208 normal films were added to improve the estimation of the specificity of the method. The second set was part of the new DDSM data set from the University of South Florida. A total of 193 cases (772 films) with 372 annotated malignancies was used. The measure of suspiciousness that was computed using the image characteristics was successful in discriminating tumours from false positive detections. Approximately 75% of all cancers were detected in at least one view at a specificity level of 0.1 false positive per image.

So‐called interval cancers of the breast: Pathologic and radiologic analysis of sixty‐four cases

Within a population‐based breast cancer screening program, 209 cancers were detected by regular mammographic screening. Additionally, 66 cancers were discovered between two consecutive screenings after one, two, or three negative screening examinations (interval cancers). The study group consisted of 25,920 women who have been participating since 1975 in a breast cancer screening program in Nijmegen, the Netherlands. In this program, single view mammography (lateromedial projection) was administered as the sole screening examination every two years. Physical examination was not part of the screening program. All previous histologic and radiologic material from 64 of those “interval” patients was available and was reviewed. In 19 of the 64 patients, direct or indirect signs of tumor were seen on the previous screening mammogram on review (observers error). In four cases, the site of the tumor lay outside the imaging field (technical error). In 41 cases, no signs of tumor could be seen on the mammograms even on review. By calculated tumor doubling times, 20 of these 41 cases were probably too small to be detected at the last screening (“real” interval cancers). However, 21 cases were probably large enough but were somehow masked from radiologic detection. The main reasons for this “masking” proved to be: 1) dense breast, 2) poorly outlined tumor mass of diffuse infiltrative type, mainly invasive lobular carcinomas, and 3) intraductal localization. The authors suggest that women with dense breasts be screened more frequently, using more views and modalities and with broader criteria for advising surgical biopsy. They also note that in general the two‐year interval between screenings is probably longer than the optimal interval.