Introduction

Recently, the relationship between ADC value and the pathological feature composing of simple tissue component is becoming clear,^1,2 in which the cell size plays an important role for generating contrast between tumor and surrounding tissue. However, the correspondence between ADC values and tumors with complicated histology, such as musculoskeletal tumors, is still unclear, and the investigational approach has not established yet. Weighted diffusion subtraction (WDS) was originally developed to eliminate T2 shine-through ambiguity in diffusion MRI evaluations³. The purpose of this study was to apply WDS technique on synovial sarcomas and to test its ability to evaluate cellularity of the tissue.

Methods

WDS is a subtraction technique, using a pixel-by-pixel calculation: $$ S(x, y) = S_0(x, y) \exp(- \Delta b D_{thres}) - S_1(x, y) \\ = S_0(x, y) \exp(- \Delta b D_{thres}) - S_0(x, y) \exp(- \Delta b ADC(x, y)), $$ where S is the signal intensity of the new image, S₀ is that of the image obtained with the low b value, and S₁ is that of the image acquired with the high b value. D_thres is the threshold value of the diffusion coefficient at which the sign of the signal of the new image is inverted, and Δb is the difference between the b values of the two diffusion-weighted images. The subjects were DWI of synovial sarcoma in three patients. The histology was proven on the specimen resected in the operation performed within 1 month after the MRI examination. The tumor locations were the hip, ankle and forearm. MR images were acquired on a 1.5T scanner (Signa HD, GE Healthcare, Waukesha, WI, USA) using body coil or quadrature foot coil as a receiver. Following routine MRI including T1WI and T2WI, EPI-based DWI (b=0 and 1000 s/mm²) was obtained with following parameters: TR/TE, 4200-5000/77-82 ms; slice thickness/spacing, 5-7/7-11 mm; matrices 128x160 or 96x128. The pixel size was ranged between 0.7 mm and 1.6 mm, depending the diameter of the subject. The image data was transferred to iMac and WDS was calculated using in-house software. WDS was obtained with the thresholds of D_thres (1.0, 1.5, 2.0, 2.5, 3.0 mm²/s). The images of WDS were visually compared to the histological images stained with hematoxylin-eosin.

Results

Fig.1 shows a representative WDS image of synovial sarcoma (D_thres = 2.0 mm²/s). The signal intensity of the tumor is heterogeneous, and the hypointense area decreases with increasing number of D_thres. The contrast between the tumor and the surrounding tissue was insufficient at D_thres of 1.0. At D_thres of 3.0, WDS clearly delineated the tumor; however, the anatomical information was lost on the image. In this study, D_thres of 2.0 was assumed to be the best value for depicting the tumor with keeping anatomical background. The hypointense on WDS with D_thres of 2.0 was well-matched at the area existing tumor cells with high cellularity on the histological specimen.

Discussion

WDS technique was originally developed to avoid T2 shine-though ambiguity. Tumor tissues were assumed to have less than 1.0 ADC values, resulting in negative pixel values in the calculated WDS images. However, in synovial sarcomas, the tumor tissue is extremely heterogeneous and ADC values are often larger than 1.0, since various components are contained within a single pixel. Using WDS with a D_thres larger than 1.0, it is possible to visualize cell component with small ADC sparsely distributed in a heterogeneous tissue. Simply stated, in the WDS process signals of tissues with ADC values close to the D_thres are suppressed, and tissues slightly smaller and larger ADC values are visualized with negative and positive signals respectively. By changing the D_thres value, it is possible to change the focus of visualisation to different range of ADC values. This is by no means a quantitative technique, but seems to be a valuable visualization tool.

Acknowledgements

No acknowledgement found.