Ductal carcinoma in situ (DCIS) of the breast, alone or in the presence of an invasive tumor (invasive ductal carcinoma [IDC]), presents a problem in its detection and treatment. DCIS represents a number of biologically different processes that exhibit variable frequencies of occult invasion and variable risks for local recurrence, after attempts at excision biopsy or lumpectomy. The term 'extensive intraductal component' or 'extensive intraductal carcinoma' (EIC) is defined as DCIS within and around an invasive tumor, comprising at least 25% of the neoplasm. Of all IDC cases, 15–30% are considered to have EIC 1. Tumors that are predominantly DCIS and appear with focal invasion are also classified as EIC.

DCIS not within the confines of the invasive tumor, but frequently in the adjacent breast tissue, makes breast-conserving surgical treatment less safe with regard to the recurrence rate. The great majority of recurrences are observed at or near the site of the primary tumor. On pathological and radiological evaluation of mastectomy specimens, some studies 1 have shown that EIC tumors commonly had DCIS not only in their immediate vicinity, but also as far as 4–6 cm from the invasive tumor. The risk of local recurrence after a breast-conserving procedure can be evaluated according to the histological type, the extent of the lesion and the state of the surgical margins. It has been reported that the local recurrence rate after a limited resection of the tumor, followed by radiotherapy, is significantly greater for EIC-positive patients than for EIC-negative patients 23. If the real extent of EIC was known preoperatively, it might be possible to safely limit the breast resection, since an initial wide excision would constitute overtreatment for the majority of patients not having a substantial amount of DCIS outside the invasive tumor. The fact that younger patients are more likely than older ones to have EIC, and that the association between EIC and breast cancer recurrence is stronger in premenopausal women than in postmenopausal women, renders the detection of such lesions imperative 1.

Mammographic evidence of microcalcifications with a certain form and distribution is presently the only method of detecting the presence of DCIS. The failure of mammography to recognize DCIS or EIC in radiologically dense breasts in young populations raises a significant problem.

Scintimammography with technetium-99m 2-methoxy isobutyl isonitrile (^99mTc-Sestamibi [^99mTc-MIBI]) has been used to detect primary breast cancer 45678. It concentrates in cancer cells by an energy-requiring transport mechanism and a transmembrane electronegative potential, in addition to nonspecific mechanisms, and it is stored within the mitochondria. Only a few reports 910 have discussed the efficacy of ^99mTc-Sestamibi in detecting DCIS.

Technetium-99m pentavalent dimercaptosuccinic acid (^99mTc-(V)DMSA) is a tumor-seeking agent already known for its efficacy in detecting medullary thyroid carcinoma. Its mechanism of accumulation has been thought to be related to the structural similarity between the (V)DMSA core and the phosphate anion (PO₄ ^3-), which is avidly taken up by some cancer cells 1112. Denoyer and colleagues recently reported that the radiotracer uptake is specifically mediated by NaPi cotransporter type III in cancer cells, similar to phosphate ions, which enter cells via NaPi cotransporters 13, while other reports suggest that its uptake in tumors is related to glucose metabolism-mediated acidosis 14. ^99mTc-(V)DMSA has been tried out in identifying invasive and preinvasive breast cancer 151617.

The present study retrospectively assessed the capability of ^99mTc-(V)DMSA and ^99mTc-Sestamibi to image DCIS and EIC. The study also investigated whether this was related to the presence of suspicious microcalcifications on mammography, stromal reaction, lymphocytic infiltration and immunohistochemical parameters such as the cell proliferation index (Ki-67), c-erbB-2, and p53 expression.

Breast cancer was histologically confirmed in 46/102 women. Of these women, 26/46 had invasive carcinomas, mainly ductal (IDC) and lobular (invasive lobular carcinoma [ILC]). This was found to be the prominent histological lesion in this group of IDC patients, and these patients were therefore classified as EIC-negative. The other 20/46 cases were diagnosed as having carcinoma in situ. Of these, 18/20 presented DCIS: in four cases (patients 9, 12, 17, and 18; see Table 2) DCIS was the sole histological finding, while the majority (14 patients) had invasive carcinoma (IDC) associated with extensive DCIS clearly outside the confines of the invasive tumor, and thus these patients were characterized as EIC-positive. Both the remaining 2/20 cases had extensive lobular carcinoma in situ (LCIS), associated with an invasive component (ILC).

EIC was found more frequently in younger women (mean age ± SD, 57.1 ± 13.7 years for EIC-positive patients versus 65.8 ± 8.0 years for EIC-negative patients; P < 0.05). The size of the invasive tumors (IDC/ILC) associated with extensive DCIS/LCIS ranged from 0.7 to 6.0 cm (mean ± SD, 2.75 ± 1.50 cm), and the size of in situ carcinomas ranged from 0.8 to 9.0 cm (mean ± SD, 4.92 ± 1.93 cm). The basic data of these 20 DCIS/LCIS cases are presented on an individual basis in Table 2.

Benign breast lesions were found in the remaining 56/102 women. Among them, epithelial hyperplasia (usual type hyperplasia and/or atypical type hyperplasia) was present in 14/56 patients, with or without fibrosis, adenosis and ductal dilatation.

All scintimammograms were characterized as positive or negative for in situ carcinoma, with reference to the diagnostic criteria set for the detection of DCIS/LCIS (i.e. presence of the pattern of diffusely increased tracer uptake, regardless of the presence of any focal increased activity). Any study displaying only focal uptake was therefore a (true or false) negative for the in situ criterion (Fig. 1). Comparative statistics (sensitivity, specificity, accuracy, positive predictive value and negative predictive value) for each radiotracer, in the total patient groups studied with each one and in the subgroup studied with both (head-to-head study), are presented in Table 3. The interobserver agreement rate between the two physicians assessing the breast scans was 92%.

Of the 20/46 breast cancer patients with DCIS/LCIS, 19/20 underwent ^99mTc-(V)DMSA and 13/20 underwent ^99mTc-Sestamibi scintimammography (12/20 were part of the head-to-head subgroup of 45 patients that underwent both). Locally diffuse, heterogeneous (patchy), poorly circumscribed increased ^99mTc-(V)DMSA accumulation, sometimes covering and surrounding focal increased accumulation (if present), was noticed in 16/17 of DCIS cases (Figs 2,3,4) and in 2/2 LCIS cases (Fig. 5), a total of 18/19 (95%) patients. The one lesion that was not detected had a size of 0.8 cm (patient 9). A similar pattern of distribution was found with ^99mTc-Sestamibi, but only in 6/13 (46%) patients. Among those 12/20 DCIS/LCIS patients that underwent both examinations, this pattern of tracer distribution was noticed in 11/12 patients (92%) with ^99mTc-(V)DMSA and in 5/12 patients (42%) with ^99mTc-Sestamibi (Tables 3 and 4).

In women with histologically confirmed usual type hyperplasia and/or atypical type hyperplasia, a similar pattern of intense diffuse, not patchy, but more homogeneous, ^99mTc-(V)DMSA distribution was observed in 10/14 patients (71%). Diffuse accumulation of ^99mTc-Sestamibi in sites of epithelial hyperplasia was observed in 6/14 patients (43%), yet it was less intense in comparison with ^99mTc-(V)DMSA (Fig. 6). In the absence of epithelial hyperplasia, the benign histological lesions of fibrosis, adenosis and ductal dilatation alone did not accumulate either of the two radiotracers. These results are also presented in Table 4.

In the series studied, the T/B ratio for ^99mTc-(V)DMSA uptake in the sites of diffuse accumulation pattern demonstrated a tendency to increase over time (Figs 2,3,4,5,6). Semiquantitative analysis (T/B ratio) performed separately in DCIS and in IDC revealed a statistically significant increase in late images only of the diffuse ^99mTc-(V)DMSA accumulation in DCIS (mean ± SD, 1.27 ± 0.22 at 10 min versus 1.76 ± 0.25 at 60 min; P < 0.01). On the contrary, T/B ratios for the focal ^99mTc-(V)DMSA uptake in invasive tumors were not significantly increased in the late acquisitions (mean ± SD, 1.76 ± 0.28 at 10 min versus 1.94 ± 0.28 at 60 min; P > 0.05). The pattern of diffuse ^99mTc-(V)DMSA distribution observed in epithelial hyperplasia demonstrated a similar tendency to increase over time. The T/B ratios for ^99mTc-Sestamibi did not demonstrate significant variability with time, either in DCIS or in IDC (mean ± SD for DCIS, 1.34 ± 0.29 at 10 min versus 1.33 ± 0.33 at 60 min; P > 0.05; and mean ± SD for IDC, 2.06 ± 0.97 at 10 min versus 1.98 ± 0.84 at 60 min; P > 0.05). Similarly, the diffuse ^99mTc-Sestamibi distribution observed in epithelial hyperplasia did not appear to increase significantly over time. Patients with LCIS and associated ILC gave parallel results, but due to the limited number of cases (n = 2) they could not be statistically evaluated.

Linear regression univariate analysis revealed no statistically significant correlation between the ^99mTc-(V)DMSA T/B ratio at 60 min and tumor size (r = 0.28, P > 0.1).

Mammography depicted suspicious microcalcifications in 10/20 (50%) patients with DCIS/LCIS (Table 2). The diffuse uptake of ^99mTc-(V)DMSA in DCIS/LCIS at 60 min was significantly higher in patients with suspicious microcalcifications on mammography as compared with patients without (mean ± SD T/B ratio, 1.81 ± 0.05 versus 1.4 ± 0.07, respectively; P = 0.003). ^99mTc-Sestamibi diffuse uptake at 60 min was not significantly different between the patient groups with and without microcalcifications (mean ± SD T/B ratio, 1.49 ± 0.19 versus 1.31 ± 0.09, respectively; P = 0.49).

An intense stromal reaction was found in 1/20 patient, a moderate stromal reaction in 4/20 patients, and a mild stromal reaction in 3/20 patients with DCIS/LCIS; a total of 8/20 patients (40%). Microcalcifications combined with stromal reaction appeared in 6/20 cases. Mild or moderate lymphocytic infiltration was noted in 4/20 patients (20%). Pure comedo was the most prominent type (10/18) among DCIS (Table 2). No statistically significant difference was found in ^99mTc-(V)DMSA or ^99mTc-Sestamibi uptake between patients with stromal reaction and/or lymphocytic infiltration and those patients without.

With reference to the studied immunohistochemical parameters in DCIS/LCIS patients, the Ki-67 value ranged from 10% to 55% (mean ± SD, 33.4 ± 13.4%), p53 ranged from 0% to 27% (mean ± SD, 6.4 ± 8.8%) and c-erbB-2 ranged from 0% to 80% (mean ± SD, 21.7 ± 26.4%) (Table 2).

The uptake of ^99mTc-(V)DMSA was significantly higher in patients with Ki-67 overexpression (≥ 40%) as compared with patients with Ki-67 levels < 40% (mean ± SD T/B ratio at 60 min, 1.86 ± 0.027 versus 1.52 ± 0.047, respectively; P = 0.025]. No significant difference was found in ^99mTc-Sestamibi scans in relation to Ki-67 levels ≥ 40% and < 40% (mean ± SD T/B ratio at 60 min, 1.51 ± 0.27 versus 1.24 ± 0.08, respectively; P = 0.46).

Overexpression of c-erbB-2 (≥ 10%) was associated with higher ^99mTc-(V)DMSA uptake in DCIS/LCIS (mean ± SD T/B ratio at 60 min, 1.89 ± 0.01 versus 1.52 ± 0.09 for values ≥ 10% and < 10%, respectively; P = 0.004). ^99mTc-Sestamibi uptake was not significantly different between women with c-erbB-2 values ≥ 10% and < 10% (mean ± SD T/B ratio at 60 min, 1.43 ± 0.06 versus 1.3 ± 0.16, respectively; P = 0.64).

Diffuse uptake of both ^99mTc-(V)DMSA and ^99mTc-Sestamibi was not significantly different with regard to p53 values < 10% and ≥ 10% (mean ± SD T/B ratio at 60 min, 1.71 ± 0.068 versus 1.72 ± 0.06, respectively, for ^99mTc-(V)DMSA; P = 0.93; and 1.60 ± 0.11 versus 1.29 ± 0.17, respectively, for ^99mTc-Sestamibi; P = 0.78).

The c-erbB-2 levels were significantly higher in DCIS cases with coexistent IDC (EIC-positive) than in EIC-negative patients (P < 0.05), whereas p53 expression did not demonstrate such a statistically significant difference.

In vitro autoradiography of the breast tumor of one patient with DCIS revealed a very good correlation between the regional distribution of ^99mTc-(V)DMSA radioactivity and the distribution of DCIS cell clumps, as observed after hematoxylin & eosin staining of the same sections (Fig. 7).

In situ breast carcinoma tumors are very difficult to identify, and the only available method to date for their detection is the evaluation of the shape and distribution of microcalcifications on mammography, with all the limitations inherent in this method (dense breasts, scars, implants, etc.).

Scintigraphic imaging of EIC, DCIS or epithelial hyperplasia with ^99mTc-(V)DMSA has already been reported 1516. Few reports have been published concerning the ability of ^99mTc-Sestamibi not actually to detect DCIS, but rather to improve invasive breast cancer detection in the presence of DCIS 910, and these studies seem to be contradictory. Howarth and colleagues 10 reported no change in sensitivity in the presence or absence of carcinoma in situ, and furthermore that there is no relationship between ^99mTc-Sestamibi uptake and DCIS, except in the presence of an invasive tumor. Several other authors, however, have suggested that diffuse ^99mTc-Sestamibi uptake in benign breast disorders was associated with proliferative changes 152223, demonstrating an increased risk of developing into breast cancer, while other workers described hormonal influence to be the cause of diffuse breast uptake on scintimammography 24. In the present study, hormonal influence or the menstrual cycle could not be responsible for this kind of tracer distribution, since all but two women were postmenopausal and they were not on any hormone replacement therapy.

In the group of patients with breast malignancy, the detected invasive tumors (IDC/ILC) presented mostly with focal increased uptake of both radiotracers. On the contrary, the pattern of diffuse heterogeneous (patchy) uptake was mainly observed in cases with DCIS, independently of the presence of an invasive tumor. Among the 26 invasive tumors not associated with extensive amounts of in situ carcinoma, only a couple of cases gave false positive results (for DCIS) with ^99mTc-(V)DMSA (Table 4): one consisted of multiple microscopic foci of IDC, and the other case presented IDC with intense lymphocytic infiltration.

LCIS was found in two cases, and due to this limited number specific scintigraphic results could not be statistically evaluated. They could have been considered as false positive findings for DCIS, yet their scintigraphic appearance was the same; they both revealed a similar diffuse inhomogeneous pattern of ^99mTc-(V)DMSA uptake, apart from the focal uptake corresponding to the coexistent ILC. They could therefore not be separated from DCIS, except after histological examination of the biopsied tissue, and thus they are studied together in this series. If this diffuse uptake was attributable to the lobular invasive component, known for its peculiar pattern of growth and local invasion, then it should have been clearly visible from the early (10 min) images, simultaneously with the focal uptake. Yet the extensive diffuse ^99mTc-(V)DMSA uptake was prominent in the late (60 min) image, whereas the focal tracer uptake was also visible in the early image (Fig. 4). ^99mTc-Sestamibi, known for its efficacy in imaging invasive carcinoma, in this case managed to image only the focal uptake in both early and late images.

The diffuse pattern of radiotracer uptake (not patchy, but more homogeneous) was also noticed in some cases of epithelial hyperplasia, more prominently with ^99mTc-(V)DMSA. Epithelial hyperplasia was therefore the major source of false positive results. The discrimination between atypical epithelial hyperplasia and in situ carcinoma is certainly not easy, even on frozen section. Yet any false positive scan attributable to hyperplasia should not be considered a major disadvantage, since it may lead to a search but not to an incorrect final decision concerning excision. This is because scintigraphic data (in addition to suspicious microcalcifications, if present) can serve as a guide and provide the surgeon with useful information preoperatively, in order to decide the extent of the search and to facilitate a guided biopsy, rather than directly determine whether an excision has to be performed.

Increased cell proliferation activity (Ki-67 ≥ 40%) was found to be significantly correlated with the diffuse ^99mTc-(V)DMSA uptake in breast carcinoma in situ. Papantoniou and colleagues 25 recently reported this factor to be independently correlated with focal ^99mTc-(V)DMSA accumulation in invasive breast cancer, while their preliminary reports 26 revealed that cases of usual type hyperplasia imaged with ^99mTc-(V)DMSA tend to be related to elevated Ki-67 values and are therefore at risk of developing breast malignancy, according to the literature 2728293031. Cutrone and colleagues claimed that increased proliferative activity is found not only in invasive cancers, but also in precancerous lesions 27.

Shaaban and colleagues 28, along with several other reports 2930, suggested that benign breast lesions such as epithelial hyperplasia, when associated with increased levels of Ki-67 and estrogen receptors type A, define a subset of hyperplastic lesions with high risk of subsequent breast cancer development. Such patients could be selected for prophylactic antiestrogen therapy to diminish the proliferative activity 2831. The observed tendency of the ^99mTc-(V)DMSA diffuse uptake pattern to increase over time (being more notable on the delayed 60-min images), as well as its relation to elevated proliferative cellular activity (Ki-67), could be considered indicative of the existence of a distinct primary pathway for the mechanism of its accumulation in tumor tissue. This is considered to reflect the tracer's uptake in structures that participate in mitotic activity of cancerous and precancerous cell populations 25. The possibility for this finding to be due to locally increased vascularity, angiogenesis or vessel permeability is therefore highly unlikely; if that was the case, then this diffuse pattern would have been equally clearly visible in the early 10-min images.

On the contrary, diffuse ^99mTc-Sestamibi distribution was not found to be significantly correlated with Ki-67 expression. Palmedo and colleagues 32 reported that the ^99mTc-Sestamibi concentration was one-half that of ^99mTc-(V)DMSA in an experimental model of rats bearing poorly differentiated breast tumours. Cutrone and colleagues found a moderate correlation between ^99mTc-Sestamibi uptake and the degree of cellular proliferation 27, while Cwikla and colleagues did not 33. Angiogenesis and an oxidative metabolism seem to be favorable factors for ^99mTc-Sestamibi tumor uptake, rather than proliferative activity 2529. This could provide an explanation for the lower sensitivity of this radiotracer, as compared with ^99mTc-(V)DMSA, in imaging such diffuse lesions.

Overexpression of c-erbB-2 (≥ 10%) is found in almost 60% of cases of high-grade comedo-type DCIS, in 10–40% of IDC and in only a few cases of ILC 3435. An increased c-erbB-2 level was a factor significantly correlated with diffuse ^99mTc-(V)DMSA uptake (but not with ^99mTc-Sestamibi uptake). This appears reasonable, since oncoprotein overexpression is usually found in aggressive in situ carcinomas with increased levels of cell proliferative activity.

The measurement of Ki-67 and c-erbB-2 on histological specimens is not related to the size of the tumor. In the present series, the T/B ratio for ^99mTc-(V)DMSA in the late image was not found to be significantly correlated with tumor size. It is therefore considered that the correlation found for the ^99mTc-(V)DMSA T/B ratio to Ki-67 and/or c-erbB-2 expression is independent of tumor size.

p53 overexpression is known to reflect tumor aggressiveness and a decreased disease-free interval following therapy. An intriguing finding requiring further investigation, however, is that p53 overexpression, unlike that of Ki-67 and c-erbB-2, was not significantly correlated with this pattern of uptake of either tracer in the present series.

DCIS of the comedo type is usually accompanied by a desmoplastic stromal reaction with pronounced neovascularization. Elastosis is also more common in DCIS 36. A variable inflammatory infiltrate is present in the periductal stroma. This consists of lymphocytes and histiocytes, in amounts ranging from sparse to abundant 37. The suggestion that stromal reaction and/or lymphocytic infiltration could be considered mediators for the diffuse radiotracer uptake was not confirmed in the studied population.

Suspicious mammographic microcalcifications were observed in only one-half of the DCIS/LCIS cases, whereas ^99mTc-(V)DMSA revealed a diffuse uptake pattern in the vast majority (18/19 patients). Nevertheless, diffuse uptake for ^99mTc-(V)DMSA (but not for ^99mTc-Sestamibi) was found to be significantly increased in women with suspicious microcalcifications as compared with those without. A possible explanation for this could be that clustered microcalcifications are usually found in more aggressive types of DCIS (such as comedo type), which tend to be related to increased Ki-67 levels. The majority of DCIS found in this series were of comedo type. Although it has been reported for the comedo type that suspicious microcalcifications approximately correspond to the histologically confirmed size, their extent in the present study was substantially smaller in comparison with the spread of diffuse ^99mTc-(V)DMSA uptake, which in turn was found to correlate very well with the histologically confirmed size of DCIS. These findings imply that calcium deposits may not represent a ^99mTc-(V)DMSA uptake mechanism, as was previously presumed 16.

Examining scintimammograms in combination with mammography could help limit false positive results. In the presence of suspicious microcalcifications, a diffuse ^99mTc-(V)DMSA uptake could be considered more suggestive of in situ carcinoma. In the absence of microcalcifications, however, a diffuse ^99mTc-(V)DMSA uptake could imply the possibility of a false positive study, probably due to epithelial hyperplasia.

In vitro ^99mTc-(V)DMSA autoradiography of one tumor specimen demonstrated a very good correlation between radioactivity uptake and the distribution of foci of DCIS cells. This is considered a strong indication for the tracer localization in these cell clumps. Autoradiography nevertheless needs to be performed in a larger series of tumor specimens to verify the findings.

^99mTc-(V)DMSA displayed an excellent sensitivity and negative predictive value in detecting DCIS/LCIS, especially in cases associated with increased cell proliferation (Ki-67) and c-erbB-2 overexpression, independent of the presence of suspicious microcalcifications. Its relatively lower specificity and positive predictive value, caused by false positive scans mainly attributable to epithelial hyperplasia, should not be considered a major disadvantage, since to date there is no diagnostic technique other than suspicious mammographic microcalcifications to define high-risk breast areas that should be biopsied. ^99mTc-Sestamibi appeared to be less sensitive, yet it demonstrated a very good specificity and could also be useful in preinvasive lesion imaging.

In our opinion, any diffuse tracer accumulation, either as an isolated finding or extending beyond the margins of a focal, well-circumscribed accumulation, could be considered to probably correspond with highly proliferating tissue and to possibly represent DCIS/LCIS or epithelial hyperplasia. Since the difference between in situ carcinoma and atypical type hyperplasia cannot always be discriminated, even histologically, we note the potential usefulness of scintimammography in imaging these lesions and making them accessible to biopsy. There is, however, a need for a larger series of patients to verify these preliminary observations, as well as prospective studies to estimate its reliability in affecting treatment decisions.

DCIS = ductal carcinoma in situ; EIC = extensive intraductal carcinoma; IDC = invasive ductal carcinoma; ILC = invasive lobular carcinoma; Ki-67 = cell proliferation index; LCIS = lobular carcinoma in situ; SD = standard deviation; T/B, tumor-to-background; ^99mTc-(V)DMSA = technetium-99m pentavalent dimercaptosuccinic acid; ^99mTc-Sestamibi (^99mTc-MIBI) = technetium-99m 2-methoxy isobutyl isonitrile.

The author(s) declare that they have no competing interests.

VP is the guarantor of integrity of the entire study and was responsible for study concepts, study design, literature research, data acquisition, data analysis/interpretation, statistical analysis, manuscript preparation, manuscript definition of intellectual content, manuscript editing, and manuscript revision/review. ST participated in data acquisition, data analysis/interpretation, statistical analysis, manuscript preparation, manuscript editing, and manuscript revision/review. EM, VV, MSou, LN, and AT participated in data acquisition, data analysis/interpretation, and manuscript preparation. MSot and IP participated in experimental studies (autoradiography in vitro), data acquisition, data analysis/interpretation, and manuscript preparation. DL, AL, and MM participated in clinical studies, data acquisition, data analysis/interpretation, and manuscript preparation. JK participated in data acquisition, data analysis/interpretation, statistical analysis, and manuscript preparation. CZ participated in data analysis/interpretation, manuscript preparation, manuscript definition of intellectual content, manuscript editing, and manuscript revision/review. All authors read and approved the final manuscript.