Spiculated breast lesions represent a recognised mammographic presentation and are strongly associated with malignancy. Spiculation is frequently observed in low-grade tumours and is often linked with favourable tumour characteristics, including oestrogen receptor positivity and lower histological grade. However, spicules may also be present in aggressive disease, rendering their clinical interpretation inconclusive. The biological relevance of the degree of spiculation relative to tumour mass remains uncertain. A retrospective exploratory single-centre investigation evaluated whether the spic mass ratio (SMR), defined as the ratio between tumour mass area including spiculations and tumour mass area without spicules, is associated with tumour characteristics and breast cancer-specific survival. The cohort comprised women with mammographically spiculated unilateral invasive breast cancer identified within the Malmö Diet and Cancer Study between 2004 and 2014. The analysis explored associations between SMR, clinicopathological variables and survival outcomes to determine whether this quantitative measure could stratify patients according to prognostic risk.
Cohort Selection and Imaging Assessment
The study population was derived from a prospective population-based cohort of 17,035 women. Among women diagnosed with breast cancer from 2004 until the end of 2014, digital mammograms were required for inclusion. Radiology reports at diagnosis had to describe a spiculated tumour appearance. Exclusions were applied if reassessment did not confirm spiculation, if cancer in situ was present or if bilateral breast cancer was identified. Of 680 women diagnosed with breast cancer during the study period, 259 had reports indicating spiculation. Exclusions included women without available digital mammograms, those lacking spiculated appearance on reassessment, and small numbers with bilateral or non-invasive disease. The final cohort consisted of 161 women with unilateral invasive spiculated breast cancer. Median age at diagnosis was 68 years, ranging from 55 to 91.
Must Read: Why Breast Cancers Are Still Missed on Mammography
Breast density had previously been assessed qualitatively from diagnostic mammography reports and categorised as fat-involuted, moderately dense or dense parenchyma according to local clinical routine aligned with BI-RADS 4th edition categories. Mode of detection was classified as screening-detected or clinically detected, with clinically detected cancers including interval cancers. SMR was calculated through manual annotation on digital mammograms using a PACS area measurement tool. The tumour mass was first outlined, followed by measurement of the combined tumour and spiculation area. Measurements were based on subjective visual interpretation without predefined operational boundaries between core and spicules. Correlation between tumour diameter and tumour area measurements was high, supporting consistency in size assessment.
Associations Between SMR and Clinicopathological Variables
SMR was analysed both as a continuous variable and stratified into tertiles defined as low, moderate and high. SMR demonstrated a statistically significant positive association with age at diagnosis, indicating higher SMR values among older women. A statistically significant association was also identified between SMR and breast density. High SMR values were most common in fat-involuted breasts, whereas low SMR values were most common in dense breasts. Differences in detection mode across SMR tertiles were not statistically significant.
No statistically significant associations were observed between SMR tertiles and tumour size, oestrogen receptor status, progesterone receptor status, human epidermal growth factor receptor 2 status, Ki67 expression category, histological grade, axillary lymph node involvement or histological type. An additional comparison evaluated the 10% of women with the highest SMR values against the remainder of the cohort. In this subgroup, all tumours were oestrogen receptor positive. Among those with available Ki67 data, most had low or intermediate Ki67 expression compared with the rest of the population. Only one woman in the high-SMR subgroup had axillary lymph node involvement. None of the nine women with SMR values equal to or greater than 2 had axillary lymph node involvement. Most SMR values in the cohort were below 2.
Survival Analysis and Methodological Considerations
Breast cancer-specific survival was analysed using Kaplan–Meier curves and Cox regression models. Follow-up was updated to December 31, 2018, with a median follow-up time of 8.45 years for the full cohort and 8.75 years among women who did not die from breast cancer. During follow-up, 18 women died from breast cancer and 33 died from other causes. No statistically significant difference in breast cancer-specific survival was observed between women with moderate or high SMR and those with low SMR. Cox regression analyses indicated a non-significant increase in hazard ratios for breast cancer death among women with moderate and high SMR compared with low SMR. These findings remained non-significant after adjustment conducted individually for age at diagnosis, breast density and tumour size. Analyses restricted to women with available tumour size data yielded similar results.
The analysis identified two statistically significant associations, namely between SMR and age and between SMR and breast density. No significant association was observed between SMR and axillary lymph node involvement or survival. The observation that none of the cases with SMR values equal to or greater than 2 had axillary lymph node involvement was based on a small number of women and was described as exploratory. Limitations included the relatively small sample size, the limited number of breast cancer-specific deaths and reliance on subjective delineation of tumour core and spicules without formal inter-reader agreement assessment. Stratification into tertiles may have limited sensitivity to detect more complex associations. Case identification was based on radiology reports describing spiculation, and breast density was assessed qualitatively in accordance with clinical routine.
The spic mass ratio, reflecting the degree of spiculation relative to tumour mass on mammography, was associated with higher age at diagnosis and with fat-involuted breast density. No statistically significant association was observed between SMR and breast cancer-specific survival or axillary lymph node involvement. Very high SMR values were observed in a small subgroup without nodal metastases, but the limited number of cases restricts interpretation. Larger investigations are required to clarify whether extensive spiculation measured by SMR carries independent prognostic significance in mammographically spiculated invasive breast cancer.
Source: Insights into Imaging
Image Credit: iStock
References:
Sturesdotter L, Sartor H, Kristensson H et al. (2026) The potential association between degree of mammographic spiculation and prognosis. Insights Imaging; 17, 29.
