Correlation of breast Magnetic Resonance Imaging with histopathology findings of suspected breast malignancy - The Nairobi experience.
PDF version





Breast cancer is the second most common cancer among women. In the United Kingdom, 40,000 cases are diagnosed annually with 80% being greater than fifty years (1).  About 17% of invasive breast cancers occur in women in their forty’s. Imaging is essential for accurate diagnosis and the early detection of breast cancers. Due to its high sensitivity for detecting invasive malignancies and the challenges in other imaging modalities, Breast Magnetic Resonance Imaging (Breast MRI) is becoming increasingly useful in the diagnosis of breast cancer. Breast MRI has been approved by the US Food and Drug Administration (FDA) since 1991 for its use as a supplemental tool, in the diagnosis of breast cancer (2).


In Kenya, according to the Nairobi cancer registry, established in 2001, using the Cancer Incidence Report Nairobi 2000-2002, Breast cancer was the most common cancer among women accounting for 23.3% with the diagnosis made using imaging modalities like mammography and breast ultrasound and ultimately diagnosed on histopathology (3).





The main objective of the study was to determine the accuracy and sensitivity of Breast MRI in the diagnosis of suspicious breast malignancies in relation to their Histopathology findings.





This was a prospective study conducted over an 8 month period, at various imaging centers in Nairobi, Kenya. The imaging was out carried using 1.5 Tesla Phillips and 1.5 Tesla GE Scanners, with the respective imaging protocols accordingly in various centers. The study incorporated patients with breast masses that had a high index of suspicion of malignancy. A specifically designed data collection form was used to record demographic details of the patients, the clinical diagnosis, the BMRI findings and the histopathologic findings.





A total of 94 adult patients presenting with suspected breast malignancy were collected from four imaging centers in Nairobi for this study: KNH, The Nairobi hospital, AKUH and Plaza Imaging Limited were analyzed. Initial analysis showed no differences in the demographic characteristics and clinical presentation of patients thus allowing a pooled analysis of data from all the sites to be conducted.


The mean and median age of presentation of patients with suspected breast cancer was 54.1 and 54 years respectively .The most commonly reported age group was 45 to 55 years; 86 patients were female, with a male-to-female ratio of 1:9. Majority of the patients were Africans. The Left breast was most frequently involved accounting for 50 cases, while 49 patients presented with a single clinical sign the most common being a single breast lump or mass.


On the imaging sequences, a hypointense signal was obtained in a similar number of lesions on both TIW and T2W images with a variation seen in 26 and 12 patients demonstrating an iso- and hyperintense signal on T2WI respectively. The predominant enhancement pattern seen was strong enhancement visualized in 60 patients, while for the enhancement graphs; the type III (washout) curve was the most predominant.


Invasive ductal carcinoma was the most common diagnosis which was followed by DCIS. In the Histopathology findings, invasive ductal carcinoma predominated and DCIS was diagnosed in fewer patients as compared to MRI. In the correlation between Breast MRI and Histopathology findings, 68 cases gave an overall percent agreement of 72.3%.





In this study, majority of the patients were female (M:F-1:9) with the most common presenting clinical sign being a breast lump/ mass and the left breast most frequently involved. The mean age of presentation was 54 years with an age range between 35 to 71 years. The most frequent range being 45 to 55 years. Majority of the masses (41.5%) diagnosed had irregular margins and with a strong enhancement pattern. In the kinetic curves, Type III (Washout) curves were the most frequent at a rate of 54.3%. Invasive ductal carcinoma was the most diagnosed cancer in both breast MRI and histopathology. In the correlation between Breast MRI and Histopathology findings, out of the 94 patients recruited, 68 cases gave an overall percent agreement of 72.3%.





Breast MRI demonstrated an overall strong percent agreement of 72.3% hence its high specificity and sensitivity in the diagnosis of breast cancers. The highest sensitivity was seen in the diagnosis of DCIS (95.2%) and specificity in malignant phylloides (100%) and breast lymphoma (100%).