Katayama Ayaka, Makhlouf Shorouk, Toss Michael S, Oyama Tetsunari, Rakha Emad A
Diagnostic Pathology, Gunma University Graduate School of Medicine, Maebashi, Japan.
Academic Unit for Translational Medical Sciences, School of Medicine, University of Nottingham, Nottingham, UK.
Histopathology. 2025 Apr;86(5):813-823. doi: 10.1111/his.15390. Epub 2024 Dec 9.
Lobular carcinoma (LC) of the breast exhibits diverse morphology and clinical behaviour. The pleomorphic variant (pLC) displays distinct cytonuclear features and aggressiveness compared to the classic variant (cLC). However, diagnosing pLC remains subjective. This study aims to refine LC's cytonuclear features, focusing on pLC.
Whole slide images of 59 LCs, including both in situ (LCIS) and invasive (ILC) lesions, were analysed. Nuclear measurements, including nuclear size and variability, were scored using QuPath image analysis software. For comparison, selected features were scored in normal cells (n = 10) and pleomorphism score-matched invasive breast carcinoma (IBC) of NST type (n = 33). Additional visual assessment of the pleomorphic ILC (pILC) cohort (n = 90) was conducted for cytomorphological features characterisation.
pILC demonstrated larger nuclear area and higher nuclear variability with abundance of cytoplasm than cILC. Compared to lymphocytes, pILC demonstrated a median area ranging from 2.7 to 4.7 times larger. Cut-off values for differentiating pILC from other ILC subtypes included median nuclear area > 48.2 μm and interquartile range (IQR) > 19.4, nuclear perimeter median > 25.2 μm and IQR > 5.3 and maximum diameter > 9.1 μm and IQR > 2.2. Multivariable logistic regression confirmed these parameters as independent predictors of pILC, with the maximum diameter being the most significant (P < 0.001). Visual assessment recognised two pILC subtypes: apocrine and non-apocrine. Apocrine variant showed nuclear roundness, pale vesicular chromatin patterns and prominent nucleoli, while non-apocrine variant exhibited greater nuclear size and shape variation.
Objective nuclear measurements, combined with cytoplasmic and architectural features, provide a robust framework for diagnosing LC subtypes, improving diagnostic accuracy and reproducibility.
乳腺小叶癌（LC）表现出多样的形态和临床行为。与经典型（cLC）相比，多形性变异型（pLC）具有独特的细胞核特征和侵袭性。然而，pLC的诊断仍然具有主观性。本研究旨在细化LC的细胞核特征，重点关注pLC。
分析了59例LC的全切片图像，包括原位（LCIS）和浸润性（ILC）病变。使用QuPath图像分析软件对包括核大小和变异性在内的核测量进行评分。为作比较，在正常细胞（n = 10）和NST型多形性浸润性乳腺癌（IBC）（n = 33）中对选定特征进行评分。对多形性ILC（pILC）队列（n = 90）进行了额外的视觉评估，以表征细胞形态学特征。
与cILC相比，pILC表现出更大的核面积、更高的核变异性以及丰富的细胞质。与淋巴细胞相比，pILC的中位面积大2.7至4.7倍。将pILC与其他ILC亚型区分开来的截断值包括中位核面积> 48.2μm且四分位间距（IQR）> 19.4、核周长中位值> 25.2μm且IQR> 5.3以及最大直径> 9.1μm且IQR> 2.2。多变量逻辑回归证实这些参数是pILC的独立预测因子，其中最大直径最为显著（P < 0.001）。视觉评估识别出两种pILC亚型：大汗腺型和非大汗腺型。大汗腺变异型表现为核圆形、淡泡状染色质模式和突出的核仁，而非大汗腺变异型则表现出更大的核大小和形状变异。
客观的核测量，结合细胞质和结构特征，为诊断LC亚型提供了一个可靠的框架，提高了诊断准确性和可重复性。
