A new study published in Diagnostic Pathology reports that this pattern is common, appearing in 29 percent of biopsy-diagnosed DCIS cases, but that it is of limited clinical relevance: it is not predictive of upstaging to invasive breast carcinoma in the subsequent resection, it is associated with younger patient age, and it is poorly reproducible between observers.
Two observers evaluated a range of morphological characteristics quantitatively or semi-quantitatively, including nuclear grade, architecture, myxoid stroma, tumour-infiltrating lymphocytes, comedonecrosis, calcifications and lobular cancerization. Immunohistochemistry for estrogen receptor, progesterone receptor and HER2 was assessed in accordance with the ASCO/CAP guidelines for invasive breast carcinoma.
Multivariate analysis revealed that DCIS with lobular cancerization presented significantly more often with myxoid stroma and with HER2 non-3+ expression, occurred more frequently in patients younger than 50 years, and presented significantly less often with comedonecrosis. These associations help characterise the pattern histologically, but the study’s central question concerned its predictive value: whether the presence of lobular cancerization in a biopsy-diagnosed DCIS could signal a higher risk that invasive carcinoma would be found in the subsequent surgical resection. The answer was no.
The study’s conclusion that lobular cancerization is not associated with upstaging removes one candidate marker from consideration.
The reproducibility results were equally striking. Nine pathologists independently analysed 148 photographs of DCIS, judging the presence or absence of lobular cancerization in each. The interobserver variability was high, indicating low diagnostic reproducibility. Kappa values ranged from -0.044 to 0.715, and Krippendorff’s alpha was 0.3549.
The authors note that a precise definition of lobular cancerization has been lacking, and that its role in DCIS has been poorly described to date. The authors note that a precise definition of lobular cancerization has been lacking, and the low reproducibility observed in this study is consistent with that diagnostic ambiguity. The study’s authors conclude that lobular cancerization in biopsy-diagnosed DCIS seems of limited clinical relevance because it is age-dependent, poorly reproducible and not associated with upstaging to invasive breast carcinoma.
The study is an open-access research article published on 22 September 2026, and the version shared carries a note that it is subject to further edits and will be replaced by the final Version of Record.
In the study's multivariate analysis, lobular cancerization was associated with younger patient age, myxoid stroma, HER2 non-3+ expression and less frequent comedonecrosis, while the low interobserver reproducibility limits the reliability of assessing this pattern in practice.
Subject of Research: Histopathological evaluation and diagnostic reproducibility of lobular cancerization in biopsy-diagnosed ductal carcinoma in situ of the breast
Article Title: Lobular cancerization in biopsy-diagnosed ductal carcinoma in situ (DCIS) of the breast: a detailed histopathological evaluation and reproducibility study
Article References: Riggi, J. A. M., Galant, C., Fontanges, Q., François, A., González Antelo, A., Lambein, K., Tamboura, A., Vernaeve, H., Vassilieff, M., Fellah, L., Gerday, A., van Deurzen, C. H. M., Berlière, M., & Van Bockstal, M. R. (2026). Lobular cancerization in biopsy-diagnosed ductal carcinoma in situ (DCIS) of the breast: a detailed histopathological evaluation and reproducibility study. Diagnostic Pathology. https://doi.org/10.1186/s13000-026-01857-w
Image Credits: AI Generated
DOI: 10.1186/s13000-026-01857-w
Keywords: ductal carcinoma in situ, DCIS, lobular cancerization, breast pathology, invasive breast carcinoma, upstaging, histopathology, interobserver variability, kappa statistic, diagnostic reproducibility, immunohistochemistry, HER2
News Source: Ophelia Keating. (October 29, 2026). Lobular cancerization in DCIS shows limited clinical value and poor reproducibility. Scienmag.



