Radiol Oncol 2024; 58(1): 33-42. doi: 10.2478/raon-2024-0017 33 research article Role of diffusion-weighted imaging in response prediction and evaluation after high dose rate brachytherapy in patients with colorectal liver metastases Salma Karim1, Ricarda Seidensticker1, Max Seidensticker1,2, Jens Ricke1,2, Regina Schinner1, Karla Treitl1, Johannes Rübenthaler1,2, Maria Ingenerf1, Christine Schmid-Tannwald1,2 1 Department of Radiology, University Hospital, LMU Munich, Germany 2 ENETS Centre of Excellence, Interdisciplinary Center of Neuroendocrine Tumours of the GastroEntero-Pancreatic System at the University Hospital of Munich (GEPNET KUM), University Hospital of Munich, Munich, Germany Radiol Oncol 2024; 58(1): 33-42. Received 10 November 2023 Accepted 4 January 2024 Correspondence to: Christine Schmid-Tannwald, Department of Radiology, University Hospital, LMU Munich, Germany; Email: Christine. schmid-tannwald@med.uni-muenchen.de Maria Ingenerf and Christine Schmid-Tannwald contributed equally to this work and share last authorship Disclosure: No potential conflicts of interest were disclosed. This is an open access article distributed under the terms of the CC-BY license (https://creativecommons.org/licenses/by/4.0/). Background. The aim of the study was to assess the role of diffusion-weighted imaging (DWI) to evaluate treatment response in patients with liver metastases of colorectal cancer. Patients and methods. In this retrospective, observational cohort study, we included 19 patients with 18 respond- ing metastases (R-Mets; follow-up at least one year) and 11 non-responding metastases (NR-Mets; local tumor re- currence within one year) who were treated with high-dose-rate brachytherapy (HDR-BT) and underwent pre- and post-interventional MRI. DWI (qualitatively, mean apparent diffusion coefficient [ADCmean], ADCmin, intraindividual change of ADCmean and ADCmin) were evaluated and compared between pre-interventional MRI, first follow-up after 3 months and second follow-up at the time of the local tumor recurrence (in NR-Mets, mean: 284 ± 122 d) or after 12 months (in R-Mets, mean: 387+/-64 d). Sensitivity, specificity, positive predictive values (PPVs), and negative predictive values (NPVs) for detection of local tumor recurrence were calculated on second follow up, evaluating (1) DWI images only, and (2) DWI with Gd-enhanced T1-weighted images on hepatobiliary phase (contrast-enhanced [CE] T1-weight [T1w] hepatobiliary phase [hb]) Results. ADCmean significantly increased 3 months after HDR-BT in both groups (R-Mets: 1.48 ± 0.44 and NR-Mets: 1.49 ± 0.19 x 10-3 mm²/s, p < 0.0001 and p = 0.01), however, intraindividual change of ADCmean (175% vs.127%, p = 0.03) and ADCmin values (0.44 ± 0.24 to 0.82 ± 0.58 x 10-3 mm2/s) significantly increased only in R-Mets (p < 0.0001 and p < 0.001). ADCmin was significant higher in R-Mets compared to NR-Mets on first follow-up (p = 0.04). Sensitivity (1 vs. 0.72), specificity (0.94 vs. 0.72), PPV (0.91 vs. 0.61) and NPV (1 vs. 0.81) could be improved by combining DWI with CE T1w hb compared to DWI only. Conclusions. DW-MRI seems to be helpful in the qualitative and quantitative evaluation of treatment response after HDR-BT of colorectal metastases in the liver. Key words: liver; HDR-brachytherapy; diffusion-weighted imaging; apparent diffusion coefficient; colorectal liver me- tastases Radiol Oncol 2024; 58(1): 33-42. Karim S et al. / Diffusions-weighted imaging at HDR-rate brachytherapy in colorectal liver metastases34 Introduction Image-guided interstitial high-dose-rate brachy- therapy (HDR-BT) supported by CT or MR fluor- oscopic-guided catheter implantation and dose calculation is a relatively new percutaneous ab- lation technique. It has shown promising results with consideration to safety, local tumor control, efficiency and overall survival (OS) in patients with unresectable liver metastases.1-4 HDR-BT can be performed repeatedly as therapy for recurrent liver metastases while maintaining liver function as high irradiation doses with steep dose gradients are being precisely applied to tumor tissue assur- ing the sparing of surrounding liver parenchyma.5 High-dose-rate brachytherapy (HDR-BT)6 of unresectable liver metastases leads to post-radio- genic changes such as post-radiogenic margins and vascularization, resulting in limited ability to assess morphological images6,7 similar to other loco-regional treatment (LRT) methods like radi- ofrequency ablation (RFA)8,9 or selective internal radiotherapy (SIRT).10,11 Currently, the modified response evaluation cri- teria in solid tumors (mRECIST) is used to evaluate treatment response of Hepatocellular carcinoma (HCC) after loco-regional treatment (LRT) strate- gies, based on tumor size and contrast agent en- hancement.12,13 However, studies have shown that these criteria might be limited because post-treat- ment contrast enhancement are not exclusive char- acteristics of viable tumor and may also be seen in benign tissue as a result of inflammation or due to post-radiogenic changes. Thus, mRECIST may underestimate treatment response.14-16 However, tumor response evaluated by the RECIST 1.1 is a morphologically-based by assessing the change in the size of the tumor and do not take into account information about the intra-lesional features.17 On the other hand side colorectal liver metastases demonstrate peripheral rim enhancement on the arterial phase and appear hypointense in the por- tal venous phase with delayed phase of enhance- ment.18 If a tumor has residual rim enhancement on the post-treatment contrast-enhanced CT (CE- CT), it may have viable tumor components , as con- firmed by pathological correlation.19 Diffusion-weighted imaging (DWI) reflects mo- tion of free water molecules and allows qualitative and quantitative (on apparent diffusion coeffi- cient [ADC] map) evaluation of changes in tissue cellularity.9,20,21 Therefore, it seems promising as a complement for evaluation of treatment response. Previous studies have already demonstrated the ability of DWI to assess tumor response in the liver after RFA as well as after SIRT to monitor different anticancer therapies.10,22-26 In addition, there are al- so studies showing that DWI may be an additional tool for predicting tumor response in patients with colorectal cancer liver metastases.24,27,28 However, there is limited data reflecting the role of diffusion-weighted imaging in evaluation of tumor response in patients with liver metasta- ses treated with HDR-BT. Wybranski et al. showed that DWI is an important parameter for early pre- diction of treatment response after HDR-BT in pa- tients with colorectal liver metastases.5 However, the study had an observation interval of only three months after therapy, and did not differentiate be- tween responding and non-responding lesions.5 Therefore, the purpose of this study was to as- sess the role of diffusion-weighted imaging in re- sponse prediction and evaluation after HDR-BT in patients with colorectal liver metastases over short and long-term intervals. Patients and methods Patients This retrospective observational cohort study was approved by the local research ethics com- mittee and the need for written informed patient consent was waived. The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology STROBE guidelines.29 Consecutively selected patients with liver me- tastases of colorectal cancer who were treated by HDR-BT at our department between August 2017 and December 2018 and who underwent MRI with DWI before, three months after HDR-BT and a sec- ond follow-up MRI at time of local tumor recur- rence in nonresponding metastases (NR-Mets) and 12 months after therapy in responding metastases (R-Mets) were evaluated. Exclusion criteria were severe motion artefacts, lesion size less than 1cm, an incomplete MRI protocol, locoregional abla- tive therapy of treated metastases before and after HDR-BT. CT-guided interstitial HDR-BT Patient selection for treatment with CT-guided HDR-BT was based on a consensus decision in an interdisciplinary tumor conference. If multiple le- sions were treated by brachytherapy in a patient, all treated lesions were included in the study. The Radiol Oncol 2024; 58(1): 33-42. Karim S et al. / Diffusions-weighted imaging at HDR-rate brachytherapy in colorectal liver metastases 35 procedure was performed in one single session as described before.6,30 After analgesia and seda- tion, CT-guided brachytherapy catheters were positioned inside the tumor volume, followed by a planning CT scan. The HDR-BT in after-loading technique was then performed using a 192Ir source. After irradiation, the catheters were removed while Gelfoam was administered to seal the punc- ture tract.6,30 The applied dose was at least 15 Gy surrounding the tumor. MR imaging Standardized pretreatment and posttreatment liver MRI were performed on a 1.5 T MR system (Magnetom Avanto, Magnetom Aera Siemens Healthcare, Erlangen, Germany or Ingenia, Ingenia S, Philips Healthcare, Best, Netherlands). Liver MRI included unenhanced T1w gradient-echo (GRE) (2D Flash) sequences in- and out-of-phase, a single shot T2w sequence (HASTE), T1w 3D GRE sequences with fat suppression (VIBE) before and 20, 50, and 120 seconds (depending on circulation time) after intravenous contrast injection (Gd-EOB- DTPA; Primovist, Eovist, Bayer Schering Pharma, Germany; 25 µmol/kg body weight), a multishot T2w turbo spin echo sequence with fat saturation, diffusion-weighted sequences with b-values of 50, 400 and 800 s/mm² and, after a delay of 15 minutes, an additional T1w GRE sequence with fat satura- tion (2D FLASH) and a fat suppressed T1w VIBE 3D GRE sequence identical to those performed ear- lier. Parallel imaging with an acceleration factor of 2 was utilized for all sequences. ADC maps were automatically computed from acquired DWI-MR images including all b-values. Image analysis Standard of reference Diagnosis of the primary tumor was established by histopathology. The evaluation of treatment re- sponse was lesion-based and based on mRECIST criteria on longterm follow-up imaging in consen- sus. Treatment evaluation was based on mRECIST but also included enhancement not only in the arte- rial but also in the portalvenous phase with corre- sponding hypointensity in the hepatobiliary phase: Complete response (CR): Disappearance of any intratumoral enhancement Partial resoinse (PR): (a) A decrease of vascu- larization of at least 30% or (b) a decrease of vas- cularization of at least 30% without washout or (c) decreasing defect/ size (at least 30%) in the hepato- biliary phase Progressive disease (PD): (a) over time increas- ing size and enhancement (at least 20%) or (b) new nodular enhancement with corresponding hy- pointensity Stable disease (SD): Treated lesions were in be- tween the three categories mentioned above. R-Mets were defined as lesions (a) without vas- cularization in the sense of a disappearance of any intratumoral enhancement (CR) or (b) a decrease of vascularization of at least 30% without washout or (c) decreasing defect/ size (at least 30%) in the hepatobiliary phase (PR) 12 months after therapy. Whereas NR-Mets were defined as treated lesions with (a) persisting (SD) and over time increasing (PD) size and enhancement (at least 20%) or (b) new nodular enhancement with corresponding hypointensity (PD) in the hepatobiliary phase. Quantitative and qualitative image analysis Image analysis was performed in consensus by two radiologists. The review was conducted in three separate sessions by two radiologists in consen- sus: (1) preinterventional MRI and (2) 1st postinter- ventional MRI and (3) 2nd post-interventional MRI with 2-week interval between the review sessions. Location and size measurements The location of each metastasis was recorded, and size measurements were performed on T1- weighted postcontrast imaging in the hepatobiliary phase on the slice with the largest tumor extent, excluding post-radiogenic changes, in consensus with all other acquired sequences. Evaluation of metastases and normal liver on DWI In each session, DWI was evaluated whether (1) metastases exhibited visually restricted diffusion or (2) showed no diffusion restriction (invisible on high b-value images or T2 shine through). Mean ADC values of tumor-free hepatic paren- chyma were measured on pre- and post-interven- tional DWI-MR images by placing circular ROIs, as large as possible, in areas of normal liver paren- chyma. For ADC measurements of the metastases cir- cular regions-of-interest (ROI) were manually drawn on the slice with the largest tumor extent on diffusion-weighted images while excluding neigh- Radiol Oncol 2024; 58(1): 33-42. Karim S et al. / Diffusions-weighted imaging at HDR-rate brachytherapy in colorectal liver metastases36 boring structures or regions close to the rim of the lesion to avoid partial volume effects. Then, ROIs were transferred to the same slice of the ADC map to calculate intralesional ADC values including minimal (ADCmin), and mean (ADCmean) ADC values (below noted as 10−3 mm2/s). Evaluation of DWI for local tumor recurrence assessment Furthermore, in two additional separate sessions, a third reviewer recorded the presence of local tumor recurrence on second follow up evaluating (1) DWI images only, (2) DWI with Gd-enhanced T1-weighted images on hepatobiliary phase (con- trast-enhanced [CE] T1-weight [T1w] hepatobiliary phase [hb]) with a four points confidence scale: 1 = no local tumor recurrence, 2 = probably no local tumor recurrence, 3 = probably local tumor recur- rence, 4 = definite local tumor recurrence. On DWI, local tumor recurrence was defined as new or increasing nodular diffusion restriction over time. On CE T1w hb, local tumor recurrence was recorded if hypointense, treated metastasis ei- ther increased in size or new hypointense lesions appeared directly adjacent to the boarder of treat- ed lesion. Statistical analysis Statistical analysis was performed with IBM SPSS Statistics 22 (IBM Corporation) and SAS (version 9.4) for Windows (SAS Institute, Inc.). All ADC values and size measurements by both readers were averaged for further statistical analy- sis. Statistical significance level was set at p ≤ 0.05. For normally distributed data, such as mean and min ADC of target lesions, paired t-tests were used for comparisons between study visits (before vs. af- ter HDR-BT) and two-sample t-tests were used for comparisons between response groups (intraindi- vidual changes in responders vs. non-responders). For non-normally distributed continuous data, such as lesion size, Mann-Whitney U test and Wilcoxon test were used instead of two-sample t-test and paired t-test. For categorical data, such as diffusion restriction McNemar’s test was used for comparisons between study visits, and Fisher’s exact test or Chi-square test was used for compari- sons between response groups. Sensitivity, speci- ficity, positive predictive values (PPVs), negative predictive values (NPVs) for detection of local tu- mor recurrence were calculated by means of cross tabulation. Significance levels of sensitivity and specificity of each review session were calculated using a McNemar Test. The Wilcoxon signed rank test for nonparametric paired samples was used for comparison of multiple confidence scores. Results Patients, MR interval, tumor location and size The final study population consisted of 19 pa- tients (6 males, 13 females; mean age: 70 years, SD: 10.7) with a total of 29 treated liver metasta- ses (Figure 1). According to reference standard, 18 metastases were rated as R-Mets in 11 patients and 11 metastases as NR-Mets in 8 patients. 11 pa- tients were responders: 6 patients each had one re- sponding lesion, 3 patients each had two respond- ing lesions and 2 patients each had 3 responding metastases after treatment with brachytherapy. 8 patients were non-responders: 5 patients each had one NR-Mets, 3 patients each had two NR-Mets. There were no patients with both, responding and non-responding lesions. Baseline imaging in R-Mets was performed 11 days (± 17 days) (MRI) before therapy, 1st and 2nd follow-up imaging were acquired 93 d (± 22 days) and 378 d (± 64 days) after HDR-BT, respectively. FIGURE 1. Flow diagram for this study DWI = diffusion-weighted imaging; HDR-BT = high-dose-rate brachytherapy; NR-Mets = non-responding metastases; R-Mets = responding metastases; HDR-BT = high-dose-rate brachytherapy Radiol Oncol 2024; 58(1): 33-42. Karim S et al. / Diffusions-weighted imaging at HDR-rate brachytherapy in colorectal liver metastases 37 Baseline imaging in NR-Mets was performed 21 days (± 17 days) before therapy, and 1st and 2nd fol- low-up imaging were acquired 96 d (± 36 days) and 284 d (± 122 days) (= at time of local recurrence) after HDR-BT, respectively. 11 lesions were located in the right lobe, and 18 lesions were located in the left. The mean size of R-Mets was 2.2 ± 1.2 cm on preinterventional MRI and decreased significantly to 1.7 ± 0.9 cm on the first postinterventional im- ages (p = 0.004) and showed another significant decrease (mean size 1.0 ± 0.4 cm) on the second postinterventional MRI (p = 0.0002). The mean size of NR-Mets also significantly decreased between pre-interventional MR images and the first postin- terventional images (4.1 ± 2.2 cm and 3.3 ± 2.0 cm, respectively). However, on second follow-up, there was again a significant increase in size (mean size: 4.1 ± 2.3 cm, p = 0.02) (Table 1). The size of R-Mets was significantly smaller than the size of NR-Mets on the pre-, 1st post- and 2nd postinterventional MRI (p = 0.007. 0.001 and p < 0.001, respectively). ADC measurements ADC of normal liver Mean ADCmean of normal liver parenchyma for patients with R-Mets was 0.93 ± 0.11x 10-3 mm²/s on preinterventional images and 0.99 ± 0.16 x 10-3 mm²/s on 1st postinterventional images and 0.88 ± 0.23 x 10-3 mm²/s on 2nd follow up. Mean ADCmean of normal liver parenchyma for patients with NR- Mets was 0.92 ± 0.18 x 10-3 mm²/s on the preinter- ventional images and 0.84 ± 0.24 x 10-3 mm²/s on the 1st post-interventional images and 0.96 ± 0.24x 10-3 mm²/s on the 2nd follow up. There were neither a statistically significant change in mean ADC values of non-tumorous liver parenchyma between base- line and follow-up MRIs (p > 0.05) nor between re- sponders and non-responders. ADCmean of metastases ADCmean of R-Mets (Table 1) was 0.84 ± 0.34 x 10-3 mm²/s on preinterventional images and 1.44 ± 0.19 x 10-3 mm²/s on the 1st postinterventional images and 1.48 ± 0.44 x 10-3 mm²/s on the 2nd follow up. There was a significant increase between baseline and the 1st follow-up examination and between baseline and the 2nd follow-up (p < 0.0001) ADCmean of NR-Mets (Table 1) also increased significantly between pre- and 1st postinterven- tional MRI (ADCmean: 1.21 ± 0.34 x 10-3 mm²/s to 1.49 ± 0.35 x 10-3 mm²/s) (p = 0.01); however, there was a significant decrease of ADCmean between 1st and 2nd follow up (ADCmean on 2. follow up: 1.28 ± 0.32 x 10-3 mm²/s, p = 0.04). The intra-individual increase in ADCmean val- ues of R-Mets after HD-BRT was 175% on first fol- low-up and 187% on second follow up compared to preinterventional MRI (Figure 2). The average increase in ADCmean values of R-Mets after HD- BRT was 12% between first and second follow-up. The intra-individual increase in ADCmean val- ues of NR-Mets after HD-BRT was 127% on first follow-up and 106% on second follow-up compared to preinterventional MRI (Figure 2). In contrast to R-Mets (Figure 3), there was a decrease of 21% in TABLE 1. Quantitative and qualitative results on baseline, 1. follow up and 2. follow up after local therapy of colorectal liver metastases with brachytherapy Target lesions Responding metastases Non-responding metastases Baseline 1. follow-up 2. follow-up Baseline 1. follow-up 2. follow-up Size (cm) 2.2 +/- 1.2 1.7 +/- 0.9 1.0 +/- 0.4 4.1 +/- 2.2 3.3 +/- 2.0 4.1 +/- 2.3 ADCmean 0.84 +/- 0.34 1.44 +/- 0.19 1.48+/- 0.44 1.21 +/- 0.34 1.49 +/- 0.35 1.28 +/- 0.32 ADCmin 0.44 +/- 0.24 0.82 +/- 0.25 0.9 +/- 0.38 0.44 +/- 0.23 0.54 +/- 0.41 0.4 +/- 0.32 Visually diffusion restriction 11/18 (61.11%) 2/18 (11.11%) 0/18 (0%) 8/11 (72.38%) 4/11 (36.36%) 8/11 (72.38%) Intraindividual increase in between baseline and 1. follow-up between baseline and 2. follow up between baseline and 1.follow-up between baseline and 2. follow up ADCmean (%) 175 187 127 106 ADCmin (%) 208 281 146 115 Radiol Oncol 2024; 58(1): 33-42. Karim S et al. / Diffusions-weighted imaging at HDR-rate brachytherapy in colorectal liver metastases38 ADCmean values of NR-Mets (Figure 4) after HD- BRT between first and second follow-up. A cut-off value of a change of ADCmean less than 39% yielded a sensitivity of 0.82 (95% CI: 0.52- 0.97) and a specificity of 0.72 (95% CI: 0.49-0.88). Comparing the intra-individual change in ADCmean values between both groups, we found a significant difference between preinterventional ADCmean to ADCmean of the first and second fol- low up: (p = 0.03 and 0.01 retrospectively). There was a significant difference of ADCmean between R-Mets and NR-Mets on preintervention- al MRI (p = 0.008). ADCmin of metastases ADCmin of R-Mets (Table 1) increased significant- ly from 0.44 ± 0.24 x 10-3 mm2/s before treatment to 0.82 ± 0.25 x 10-3 mm2/s after treatment on the first follow up and to 0.9 ± 0.38 x 10-3 mm2/s on the sec- ond follow-up (p < 0.0001 and py0.001), but there was no significant increase of ADCmin between the first and second follow up (p = 0.49). In NR- Mets (Table 1) there was no significant change of ADCmin (from 0.44 ± 0.23 x 10-3 mm2/s to 0.54 ± 0.41 x 10-3 mm2/s to 0.40 ± 0.32 x 10-3 mm2/s) over time. The intra-individual increase in ADCmin val- ues of R-Mets after HD-BRT was 208% on first fol- low up and 281% on second follow up compared to preinterventional MRI. The intra-individual increase in ADCmin val- ues of NR-Mets after HD-BRT was 146% on the first follow up and 115% on the second follow up compared to preinterventional MRI. There were no significant difference between preinterventional ADCmin to ADCmin of the sec- ond follow up: (p = 0.03) but not compared to first follow up (p = 0.1). There was no significant differences of preinter- ventional ADCmin between R-Mets and NR-Mets, however ADCmin was significantly higher in R-Mets compared to NR-Mets on the first follow- up (p = 0.04). Qualitative analysis of metastases In R-Mets, there was a significant loss of diffusion restriction over time (pre- to first postinterven- tional MRI p = 0.012 and pre- to second postint- erventional p = 0.001): On preinterventional MRI, 11/18 (61.11%) R-Mets were diffusion-restricted. On the first follow-up, only 2/18 (11.11%) showed diffusion restriction and on the second follow up, no responding metastasis showed restricted diffu- sion. In contrast, 8/11 (72.73%) NR-Mets showed restricted diffusion on preinterventional, 4/11 (36.36%) on the first follow up and then again 8/11 (72.73%) showed restricted diffusion on the second follow-up. Detection of local tumor recurrence on DWI There were 11 recurrent lesions in total on the second follow-up. On DWI only, 8 of 11 NR- Mets and 13 of 18 R-Mets were correctly detected. Combining DWI with the hepatobiliary phase 11 of 11 NR-Mets and 17 of 18 R-Mets were identified. It was not differentiated whether the lesions showed enhancement in the first follow-up and therefore showed a stable disease in the short-term interval (SD) or whether the lesions showed a completely new nodular enhancement with corresponding hypointensity on the hepatobiliary phase in the second follow-up (PD), since both types of lesions were classified as NR Mets. The presence of local tumor recurrence on the second follow up evaluating DW-images only re- sulted in a sensitivity of 0.72, a specificity of 0.72, a positive predictive value (PPV) of 0.61 and a nega- tive predictive value (NPV) of 0.81. FIGURE 2. Mean apparent diffusion coefficient (ADCmean) change in responding metastases (R-Mets) and NR-Mets between preinterventional MRI and first and second follow-up, respectively ADC = apparent diffusion coefficient; NR-Mets = non-responding metastases, R-Mets = responding metastases Radiol Oncol 2024; 58(1): 33-42. Karim S et al. / Diffusions-weighted imaging at HDR-rate brachytherapy in colorectal liver metastases 39 Combining DWI with Gd-enhanced T1- weighted images on hepatobiliary phase improved diagnostic performance: Sensitivity: 1, specificity 0.94, PPV: 0.91, NPV: 1. Discussion Loco-regional treatment methods like SBRT or HDR-BT lead to post-radiation changes such as cell swelling, transudation of plasma components to the extravascular-extracellular and space of tu- mor but also cellular necrosis and changes in mi- crovasculature.5,6 DWI reflects changes in tumor cellularity and cell membrane integrity but also vascular capillary perfusion.20 In the current lit- erature quantitative and qualitative evaluation of DWI seems to be a promising tool in evaluat- ing tumor response after loco-regional treatment; to the best of our knowledge there are no studies determining the role of DWI in patients with colo- rectal metastases treated with HDR-BT to stratify responding from non-responding metastases and therefore, determining the role of DWI for predic- tion of tumor response. In the early follow-up, three months after treat- ment with HDR-BT, we found in both groups a significant increase of ADCmean. We suggest that high dose rate brachytherapy induces loss of cell membrane integrity, increased extracellular space and ultimately tumor cell lysis. However, regard- ing the intraindividual change of ADCmean, only responding lesions showed a significant increases in ADCmean which correlates with other studies FIGURE 3. R-Met in a 62-year-old female. The pre-interventional diffusion-weighted imaging (DWI) shows two diffusion-restricted liver metastases with high signal on axial diffusion-weighted (DW)-MR image b = 800 s/mm2 (A) and low signal on apparent diffusion coefficient (ADC) map (B). The pre-interventional ADCmean of the metastases were 0.83 and 0.86 x 10-3 mm²/s. In the hepatobiliary phase (C) both metastases showed a hypointense signal. After high-dose-rate brachytherapy (HDR-BT), the metastases demonstrated a hyperintense signal on the axial DW-MR image (D) and a hyperintense signal on the ADC map (E) indicating less restricted diffusion compared to the pre-interventional image. The ADCmean increased to 1.41 and 1.53 x 10-3 mm²/s. in the hepatobiliary phase (F). The lesion showed central necrosis with a peripheral post-radiogenic hypointense rim. In the second follow-up the lesions showed no restricted diffusion (G) with a further increasing ADC (H) value of 2.09 and 2.07 x 10-3 mm²/s. There was a shrinkage in size of the metastases without a new hypointense defect in the hepatobiliary phase (I). A B C D G E H F I Radiol Oncol 2024; 58(1): 33-42. Karim S et al. / Diffusions-weighted imaging at HDR-rate brachytherapy in colorectal liver metastases40 in the literature.25,31 Furthermore ADCmin signifi- cantly increased in the responding lesions indi- cating necrosis induction in the tumor periphery beyond the already pre-interventional persisting central necrosis which is well known in colorec- tal metastases.25,29-32 It seems that DWI may be also used in long-term evaluation of tumor response. We found that over time (12months) a significant increase of ADCmean and ADCmin as well as loss of diffusion restriction could only be observed in responding metastases. In contrast, non-respond- ing metastases even showed at time of local recur- rence a decrease in ADCmean and ADCmin and consecutively increasing visual diffusion restric- tion compared to the first/early follow-up. In the current literature there seems to be a disa- greement regarding the role of pretreatment ADC values in predicting tumor response, depending on treatment techniques applied: Cui et al. as well as Koh et al. demonstrated that pretreatment ADCmean values in non-responding lesions of patients with colorectal and gastric he- patic metastases treated by chemotherapy were significantly higher than those of nonresponding lesions.22,24 Before chemotherapy, the presence of necrosis (resulting in higher ADC values) may lead to less delivery of chemotherapeutic drugs to these less perfused regions.24 Similar results were found by Lahrsow et al.28: Responding colorectal liver metastases had sig- nificant lower ADC values than non-responding metastases before treatment with conventional lipiodol-based transarterial chemoembolization. In contrast, Schmeel et al. showed that ADCmean in responding hepatic metastases of colorectal ori- gin treated with 90Y-microsphere radioemboliza- tion were significantly higher than ADCmean of non-responding lesions.27 This might be attributed FIGURE 4. Non-responding metastases (NR-met) in a 56-year-old male. In preinterventional MRI, metastasis (circle) shows restricted diffusion (A+B) with an mean apparent diffusion coefficient (ADCmean) of 0.86 x 10-3 mm²/s and a hypointense pattern on the liver-specific phase (C). Three months after high-dose-rate brachytherapy (HDR-BT), the metastasis showed visually partial restricted diffusion (D+E), but, with an increasing ADCmean of 1.52 x 10-3 mm²/s and hypointensity in the hepatobiliary phase (F). After 11 months, the lesion increased in size, shows a visually an increasing diffusion restriction (K+L) at the boarder (arrow) with a persistently ADCmean value of 1.53 x 10-3 mm²/s and a new defect in the hepatobiliary phase (arrow) (I) indicating local tumor recurrence. A B C D G E H F I Radiol Oncol 2024; 58(1): 33-42. Karim S et al. / Diffusions-weighted imaging at HDR-rate brachytherapy in colorectal liver metastases 41 to the higher tumor grade and tumor aggressive- ness associated with highly diffusivity restricted tumors.27,33 We found that preinterventional ADCmean was significantly lower in responding lesions com- pared to non-responding lesions. However, this re- sult can only be seen as a marginal result and was not the central issue. In addition, as a limiting fac- tor with regard to the value of our result, it must be noted that in our group the size of the metastases differed significantly between the responders and non-responders in the baseline examination. In this study we could achieve good sensitivity and specificity in detection of local tumor recur- rence; however, combining DWI with T1-weighted images in the hepatobiliary phase increased sensi- tivity, specificity and PPV and NPVv as well. Still, it must be noted, that there is a certain bias in the evaluation of combining DWI with T1-weighted images in the hepatobiliary phase, since T1- weighted images in the hepatobiliary phase was part of the gold standard that we have defined. However, at least one lesion was scored as a false positive combining DWI with T1-weighted images in the hepatobiliary phase, i.e. NR-Met, underlin- ing that the contrast media behavior of the lesions should be part of the response assessment if pos- sible. On the other hand, DWI alone achieves good but also worse results than the contrast-enhanced- based evaluation. Similar results were found by Liu et al. in detecting residual HCC after drug- eluting bead transarterial chemoembolization us- ing DWI.34 Especially in very severe renal function impaired patients or in case of general avoidance of gadolinium exposure taking into account the frequent examinations over years in typical cases or in case of contraindication for contrast media administration DWI might be a valid alternative at imaging follow up after HDR-BT. Our study has limitations due to its retrospec- tive design, single center study and small sample size, which generally limits the conclusions to be drawn due to the lack of reproducibility. On the other hand, the evaluation was lesion-based with at least a total of 29 lesions were evaluated. Furthermore, the timing of imaging acquisition before and especially after treatment between the responder and non-responder group was not en- tirely similar. Moreover, mRECIST is limited by post-interventional changes that simulate a local recurrence. However, to overcome this limita- tion, we chose a long period after the intervention for the final evaluation. DWI is already routinely used to assess a treatment response of liver me- tastases and has been evaluated in many studies. Nevertheless, the use of DWI to assess a treatment response after brachytherapy in the liver has not yet been evaluated in short- and long-term stud- ies; the post-radiogenic changes, especially after brachytherapy, can be delineated in a circular hy- pointensity in the hepatobiliary phase around the lesion for a relatively long time and thus the as- sessment of response is difficult, as a result pro- viding the study basis. In conclusion, our results indicate that DWI- MRI may be a useful adjunct to morphologic MRI for detection of local tumor recurrence in patients with colorectal liver metastases treated with HDR- BT. References 1. Ricke J, Wust P, Stohlmann A, Beck A, Cho CH, Pech M, et al. CT-guided interstitial brachytherapy of liver malignancies alone or in combination with thermal ablation: phase I-II results of a novel technique. Int J Radiat Oncol Biol Phys 2004; 58: 1496-505. doi: 10.1016/j.ijrobp.2003.09.024 2. Ricke J, Wust P, Wieners G, Beck A, Cho CH, Seidensticker M, et al. Liver ma- lignancies: CT-guided interstitial brachytherapy in patients with unfavorable lesions for thermal ablation. J Vasc Interv Radiol 2004; 15: 1279-86. doi: 10.1097/01.Rvi.0000141343.43441.06 3. Seidensticker M, Wust P, Rühl R, Mohnike K, Pech M, Wieners G, et al. Safety margin in irradiation of colorectal liver metastases: assessment of the control dose of micrometastases. Radiat Oncol 2010; 5: 24. doi: 10.1186/1748-717x-5-24 4. Ricke J, Mohnike K, Pech M, Seidensticker M, Rühl R, Wieners G, et al. Local response and impact on survival after local ablation of liver metastases from colorectal carcinoma by computed tomography-guided high-dose-rate brachytherapy. Int J Radiat Oncol Biol Phys 2010; 78: 479-85. doi: 10.1016/j. ijrobp.2009.09.026 5. Wybranski C, Zeile M, Löwenthal D, Fischbach F, Pech M, Röhl FW, et al. Value of diffusion weighted MR imaging as an early surrogate parameter for evaluation of tumor response to high-dose-rate brachytherapy of colorectal liver metastases. Radiat Oncol 2011; 6: 43. doi: 10.1186/1748-717x-6-43 6. Bretschneider T, Ricke J, Gebauer B, Streitparth F. Image-guided high-dose- rate brachytherapy of malignancies in various inner organs − technique, indications, and perspectives. J Contemp Brachytherapy 2016; 8: 251-61. doi: 10.5114/jcb.2016.61068 7. Omari J, Drewes R, Orthmer M, Hass P, Pech M, Powerski M. Treatment of metastatic gastric adenocarcinoma with image-guided high-dose rate, inter- stitial brachytherapy as second-line or salvage therapy. Diagn IntervRradiol 2019; 25: 360-7. doi: 10.5152/dir.2019.18390 8. Chow DH, Sinn LH, Ng KK, Lam CM, Yuen J, Fan ST, et al. Radiofrequency ablation for hepatocellular carcinoma and metastatic liver tumors: a com- parative study. J Surg Oncol 2006; 94: 565-71. doi: 10.1002/jso.20674 9. Le Bihan D, Turner R, Douek P, Patronas N. Diffusion MR imaging: clini- cal applications. AJR Am J Roentgenol 1992; 159: 591-9. doi: 10.2214/ ajr.159.3.1503032 10. Dudeck O, Zeile M, Wybranski C, Schulmeister A, Fischbach F, Pech M, et al. Early prediction of anticancer effects with diffusion-weighted MR imaging in patients with colorectal liver metastases following selective internal radio- therapy. Eur Radiol 2010; 20: 2699-706. doi: 10.1007/s00330-010-1846-z 11. Welsh JS, Kennedy AS, Thomadsen B. Selective internal radiation thera- py (SIRT) for liver metastases secondary to colorectal adenocarcinoma. Int J Radiat Oncol Biol Phys 2006; 66(2 Suppl): S62-73. doi: 10.1016/j. ijrobp.2005.09.011 Radiol Oncol 2024; 58(1): 33-42. Karim S et al. / Diffusions-weighted imaging at HDR-rate brachytherapy in colorectal liver metastases42 12. Lencioni R, Llovet JM. Modified RECIST (mRECIST) assessment for hepato- cellular carcinoma. Semin Liver Dis 2010; 30: 52-60. doi: 10.1055/s-0030- 1247132 13. Kim MN, Kim BK, Han KH, Kim SU. Evolution from WHO to EASL and mRECIST for hepatocellular carcinoma: considerations for tumor response assessment. Expert Rev Gastroenterol Hepatol 2015; 9: 335-48. doi: 10.1586/17474124.2015.959929 14. Vouche M, Kulik L, Atassi R, Memon K, Hickey R, Ganger D, et al. Radiological-pathological analysis of WHO, RECIST, EASL, mRECIST and DWI: imaging analysis from a prospective randomized trial of Y90 ± Sorafenib. Hepatology 2013; 58: 1655-66. doi: 10.1002/hep.26487 15. Vossen JA, Buijs M, Kamel IR. Assessment of tumor response on MR imag- ing after locoregional therapy. Tech Vasc Interv Radiol 2006; 9: 125-32. doi: 10.1053/j.tvir.2007.02.004 16. Ludwig JM, Camacho JC, Kokabi N, Xing M, Kim HS. The role of diffusion- weighted imaging (DWI) in Locoregional therapy outcome prediction and response assessment for hepatocellular carcinoma (HCC): the new era of functional imaging biomarkers. Diagnostics 2015; 5: 546-63. doi: 10.3390/ diagnostics5040546 17. Eisenhauer EA, Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer 2009; 45: 228-47. doi: 10.1016/j.ejca.2008.10.026 18. Hamm B, Thoeni RF, Gould RG, Bernardino ME, Lüning M, Saini S, et al. Focal liver lesions: characterization with nonenhanced and dynamic con- trast material-enhanced MR imaging. Radiology 1994; 190: 417-23. doi: 10.1148/radiology.190.2.8284392 19. Ishida K, Tamura A, Kato K, Uesugi N, Osakabe M, Eizuka M, et al. Correlation between CT morphologic appearance and histologic findings in colorectal liver metastasis after preoperative chemotherapy. Abdom Radiol (NY) 2018; 43: 2991-3000. doi: 10.1007/s00261-018-1588-y 20. Koh DM, Collins DJ. Diffusion-weighted MRI in the body: applications and challenges in oncology. AJR Am J Roentgenol 2007; 188: 1622-35. doi: 10.2214/ajr.06.1403 21. Le Bihan D, Breton E, Lallemand D, Grenier P, Cabanis E, Laval-Jeantet M. MR imaging of intravoxel incoherent motions: application to diffusion and per- fusion in neurologic disorders. Radiology 1986; 161: 401-7. doi: 10.1148/ radiology.161.2.3763909 22. Cui Y, Zhang XP, Sun YS, Tang L, Shen L. Apparent diffusion coefficient: po- tential imaging biomarker for prediction and early detection of response to chemotherapy in hepatic metastases. Radiology 2008; 248: 894-900. doi: 10.1148/radiol.2483071407 23. Eccles CL, Haider EA, Haider MA, Fung S, Lockwood G, Dawson LA. Change in diffusion weighted MRI during liver cancer radiotherapy: preliminary obser- vations. Acta Oncol 2009; 48: 1034-43. doi: 10.1080/02841860903099972 24. Koh DM, Scurr E, Collins D, Kanber B, Norman A, Leach MO, et al. Predicting response of colorectal hepatic metastasis: value of pretreatment apparent diffusion coefficients. AJR Am J Roentgenol 2007; 188: 1001-8. doi: 10.2214/ ajr.06.0601 25. Marugami N, Tanaka T, Kitano S, Hirohashi S, Nishiofuku H, Takahashi A, et al. Early detection of therapeutic response to hepatic arterial infusion chemotherapy of liver metastases from colorectal cancer using diffusion- weighted MR imaging. Cardiovasc Intervent Radiol 2009; 32: 638-46. doi: 10.1007/s00270-009-9532-8 26. Schraml C, Schwenzer NF, Clasen S, Rempp HJ, Martirosian P, Claussen CD, et al. Navigator respiratory-triggered diffusion-weighted imaging in the follow-up after hepatic radiofrequency ablation-initial results. J Magn Reson Imaging 2009; 29: 1308-16. doi: 10.1002/jmri.21770 27. Schmeel FC, Simon B, Luetkens JA, Traber F, Meyer C, Schmeel LC, et al. Prognostic value of pretreatment diffusion-weighted magnetic resonance imaging for outcome prediction of colorectal cancer liver metastases un- dergoing 90Y-microsphere radioembolization. J Cancer Res Clin Oncol 2017; 143: 1531-41. doi: 10.1007/s00432-017-2395-5 28. Lahrsow M, Albrecht MH, Bickford MW, Vogl TJ. Predicting treatment re- sponse of colorectal cancer liver metastases to conventional lipiodol-based transarterial chemoembolization using diffusion-weighted MR imaging: value of pretreatment apparent diffusion coefficients (ADC) and ADC changes under therapy. Cardiovasc Intervent Radiol 2017; 40: 852-9. doi: 10.1007/s00270-017-1634-0 29. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol 2008; 61: 344-9. doi: 10.1016/j.jclinepi.2007.11.008 30. Mohnike K, Wieners G, Schwartz F, Seidensticker M, Pech M, Ruehl R, et al. Computed tomography-guided high-dose-rate brachytherapy in hepatocel- lular carcinoma: safety, efficacy, and effect on survival. Int J Radiat Oncol Biol Phys 2010; 78: 172-9. doi: 10.1016/j.ijrobp.2009.07.1700 31. Ingenerf MK, Karim H, Fink N, Ilhan H, Ricke J, Treitl KM, et al. Apparent diffu- sion coefficients (ADC) in response assessment of transarterial radioemboli- zation (TARE) for liver metastases of neuroendocrine tumors (NET): a feasibil- ity study. Acta Radiol 2022; 63: 877-88. doi: 10.1177/02841851211024004 32. Outwater E, Tomaszewski JE, Daly JM, Kressel HY. Hepatic colorectal me- tastases: correlation of MR imaging and pathologic appearance. Radiology 1991; 180: 327-32. doi: 10.1148/radiology.180.2.2068294 33. Curvo-Semedo L, Lambregts DM, Maas M, Beets GL, Caseiro-Alves F, Beets- Tan RG. Diffusion-weighted MRI in rectal cancer: apparent diffusion coef- ficient as a potential noninvasive marker of tumor aggressiveness. J Magn Reson Imaging 2012; 35: 1365-71. doi: 10.1002/jmri.23589 34. Liu Z, Fan JM, He C, Li ZF, Xu YS, Li Z, et al. Utility of diffusion weighted imaging with the quantitative apparent diffusion coefficient in diagnosing residual or recurrent hepatocellular carcinoma after transarterial chem- oembolization: a meta-analysis. Cancer Imaging 2020; 20: 3. doi: 10.1186/ s40644-019-0282-9