Diffusion weighted imaging to predict nodal status in breast cancer: A systematic review and meta‐analysis

Abstract

Diffusion weighted imaging is a functional imaging modality based upon Brownian water movement in tissues.1– 3 This movement can be quantified by apparent diffusion coefficient (ADC). According to the literature, apparent diffusion coefficient (ADC) inversely correlates with histopathological features.1,4 It was identified that benign breast lesions have significant higher ADC values compared to breast cancer (BC).5 Nodal status is one of the most important prognosis parameters in BC, which is assessed by axillary lymph node dissection.6 Our purpose was to systematically review the published literature regarding ADC values of BC in accordance with nodal status and the ADC values of axillary lymph nodes and perform a metaanalysis to establish, whether ADC values can reliably predict nodal status in BC. The literature was searched for suitable papers investigating the associations between ADC values and nodal status in BC. The metaanalysis was performed as previously described.5 The primary end point of the systematic review was association between nodal status of BC and ADC values of the primary tumor and axillary LN. In total, 23 studies were suitable for the analysis and were included into the present study. The acquired 23 studies comprised overall 1669 BC and 1423 LN. In total, 1025 nodal negative and 548 nodal positive BC were included into the analysis. The mean ADC of the primary tumor of the nodal positive BC was 0.89 × 10− 3 mm2/s [95% CI 0.84– 0.94, I2 = 91%], and 0.96 × 10− 3 mm2/s [95% CI 0.91– 1.01, I2 = 92%] of the nodal negative BC (Figure 1). In total, 876 benign and 547 malignant LN were included into the analysis. The mean ADC value of the malignant axillary LN was 0.90 × 10− 3 mm2/s [95% CI 0.80– 1.01, I2 = 98%] and 1.17 × 10− 3 mm2/s [95% CI 1.02– 1.32, I2 = 99%] of the benign LN (Figure 2). Three studies reported data regarding ADC values and LN size in large and short axis with overall 57 lymph nodes. Spearman s correlation coefficient for the large axis was r = −0.143, p = 0.29 and for the short axis was r = 0.024, p = 0.86. The present analysis showed that ADC values of the axillary LN are lower in malignant LN compared to benign LN and might aid in clinical decisionmaking. Similar findings were identified of the ADC values of primary tumors to predict nodal status in BC. Thus, the present analysis based upon a large sample size corroborates the results of previous singlecenter studies. Nodal status is one of the most important prognosis factors in BC, which leads to a worse outcome in patients with positive axillary lymph nodes compared to nodal negative patients.6 A precise definition of the status of axillary LN in BC patients is also important to select treatment modality or surgical method. Axillary LN dissection is the reference standard for evaluating nodal involvement. Several imaging modalities were used to predict LN status noninvasively. So, ultrasound was deemed to be unreliable.7 The cost expensive modality with radiation exposure, namely PETCT, achieved only a pooled sensitivity of 56% in a recent metaanalysis, whereas morphological MRI achieved a sensitivity of 55% only.7 However, both modalities reached a high specificity of 91% and 86%, respectively. Yet, there is definite need for a better diagnostic modality to predict nodal status in BC. The present analysis tried to emphasize that DWI and ADC values might be able to fill this diagnostic gap with measurement of the primary tumor itself as well as with the measurement of the lymph nodes. Interestingly, ADC values are independent of the lymph node size, which is most often used in clinical routine