1816

Quantitative intravoxel incoherent motion (IVIM) parameters derived from whole-tumor volume for predicting and assessing pathological complete response to neoadjuvant chemotherapy in locally advanced rectal cancer
Qiaoyu Xu1,2, Yanyan Xu1, Queenie Chan3, and Hongliang Sun1

1Radiology, China-Japan Friendship Hospital, Beijing, China, 2Radiology, Beijing Chao-yang Hospital, Beijing, China, 3Research center, Philips Healthcare, HongKong, China

Synopsis

Neoadjuvant chemotherapy has been widely applied in treating locally advanced rectal cancer (LARC) for effectively decreasing the local recurrence after total mesorectal excision(TME)1. Some of patients may achieve a pathological complete response (pCR). The pCR rate of LARC has been reported to be 10%–30%2-3. The imaging technique of intravoxel incoherent motion(IVIM) model, which could estimate tissue perfusion and diffusion components individually using multi-b-values was introduced to evaluate tumor treatment response4. In this study, we found that IVIM-dreived D values is a promising tool in predicting pCR status before therapy.

Synopsis

Neoadjuvant chemotherapy has been widely applied in treating locally advanced rectal cancer (LARC) for effectively decreasing the local recurrence after total mesorectal excision(TME)1. Some of patients may achieve a pathological complete response (pCR). The pCR rate of LARC has been reported to be 10%–30%2-3. The imaging technique of intravoxel incoherent motion(IVIM) model, which could estimate tissue perfusion and diffusion components individually using multi-b-values was introduced to evaluate tumor treatment response4. In this study, we found that IVIM-dreived D values is a promising tool in predicting pCR status before therapy.

Purpose

The purpose of this study was to identify whether intravoxel incoherent motion (IVIM) parameters derived from whole-tumor volume (WTV) before and after neoadjuvant chemotherapy (NACT) can accurately predict and assess pathological complete response (pCR) in patients with locally advanced rectal cancer (LARC).

Materials and Methods

The study enrolled fifty-one patients with LARC who underwent magnetic resonance imaging (MRI) on a 3-Tesla MRI scanner before and after NACT prior to surgery.Apparent diffusion coefficient (ADC),slow diffusion coefficient (D),fast diffusion coefficient (D*),and perfusion-related diffusion fraction (f) values were obtained on diffusion-weighted magnetic resonance (DW-MR) images by WTV methods and calculated by using biexponential model before and after NACT.Compared with histopathologic findings,DWI-derived ADC and IVIM-derived parameters and their percentage changes (ΔADC%,ΔD%, ΔD*%,Δf%) were compared by using independent-samples T test and Mann-Whitney U test between the pCR group and non-pCR group.Intra and interobserver variability was analyzed by calculating intraclass correlation coefficient (ICC).Diagnostic performance of IVIM parameters and their percentage changes were evaluated using receiver operating characteristic curves (ROC).

Results

Patients were assigned to pCR group (n=10) and non-pCR group (n=41) based on the histopathologic examination results following surgery.Compared with non-pCR group,the pCR group exhibited significantly lower pre-ADCmean (p=0.003)and pre-D values (p=0.024),and significantly higher post-f (p=0.002),ΔADCmean %(p=0.002),ΔD% (p=0.001),and Δf% values (p=0.017).pre-D value had the best specificity (95.12%) and accuracy (86.27%) in predicting pCR status,and ΔD% had the highest area under curve (0.832) in assessing pCR response to NACT.Intra and interobserver agreement of whole-tumor volume analysis was excellent.

Discussion

The present study demonstrated that the pCR group had significantly lower ADC and D values before NACT compared with the non-pCR group, which might help clinicians predict the pCR to NACT in LARC. The percentage changes in ADC, D, and f values before and after NACT were significantly higher in the pCR group, suggesting that ADC, D, and f values might be useful in the assessing pCR response before surgery. In addition, the percentage changes in D values before and after NACT showed the highest diagnostic performance, and D values before therapy showed the best specificity and accuracy among all parameters. The WTV measurement showed excellent reproducibility.

Conclusion

IVIM-derived D values is a promising tool in predicting pCR status before therapy, and the percentage changes of D values after therapy may be helpful in assessing pCR status without surgery. WTV analysis s shows the excellent reproducibility both inter and intraobservers.

Acknowledgements

No acknowledgement found.

References

1. Hötker AM, Garcia-Aguilar J, Gollub MJ. Multiparametric MRI of Rectal Cancer in the Assessment of Response to Therapy. Dis Colon Rectum 2014;57(6):790-799.

2. Yeo S, Kim DY, Kim TH, et al. Pathologic Complete Response of Primary Tumor Following Preoperative Chemoradiotherapy for Locally Advanced Rectal Cancer. Ann Surg 2010;252(6):998-1004. 3. Wallin U, Rothenberger D, Lowry A, Luepker R, Mellgren A. CEA - A Predictor for Pathologic Complete Response After Neoadjuvant Therapy for Rectal Cancer. Dis Colon Rectum 2013;56(7):859-868.

4. Iima M, Le Bihan D. Clinical Intravoxel Incoherent Motion and Diffusion MR Imaging: Past, Present, and Future. Radiology 2016;278(1):13-32.

Figures

Figure 1. A 61-year-old man with LARC from the pCR group (TRG 4). Images in row A are T2,DWI,ADC,D,D*,and f maps before NACT. The ADC, D, D*,and f values were 0.840× 10–3mm2/s, 0.804× 10–3mm2/s, 9.321× 10–3mm2/s, and 0.319, respectively. Images in row B are T2, DWI, ADC, D, D*, and f maps after NACT. The ADC, D, D*, and f values were 1.270× 10–3mm2/s, 1.094× 10–3mm2/s, 7.460× 10–3mm2/s, and 0.362, respectively.

Figure 2. A 75-year-old man with LARC from the non-pCR group (TRG 2). Images in row A are T2, DWI, ADC, D, D*, and f maps before NACT. The ADC, D, D*, and f values were 1.140×10–3mm2/s, 0.811×10–3mm2/s, 11.384× 10–3mm2/s, and 0.320,respectively. Images in row B are T2, DWI, ADC, D, D*, and f maps after NACT. The ADC, D, D*, and f values were 1.180× 10–3mm2/s, 0.852× 10–3mm2/s,10.539× 10–3mm2/s, and 0.313, respectively.

Figure 3. (A) ROC curve analysis of pre-ADCmean and pre-D values for predicting pCR.(B) ROC curve analysis of post-f, DADCmean, DD, and Df values for assessing pCR

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)
1816