Percentage apparent diffusion coefficient change for early prediction of response to neoadjuvant chemotherapy in breast cancer: a systematic review and meta-analysis, and a retrospective study

Introduction

Breast cancer remains a major global health burden, accounting for an estimated 2.3 million new cases and 665,684 deaths annually according to GLOBOCAN 2022 data (1). Neoadjuvant chemotherapy (NACT), defined as systemic cytotoxic therapy administered prior to surgery, has become a cornerstone of contemporary breast cancer management for dual clinical purposes (2,3). For initially inoperable disease, NACT helps achieving tumor volume reduction to make surgery possible without increasing the rates of complications (4). For operable candidates, it enhances breast conservation rates while maintaining oncological safety (5). Pathological complete response (pCR) is the ideal short-term ending of NACT, and it demonstrates strong correlation with improved long-term outcomes such as event-free survival (EFS), disease-free survival (DFS), and overall survival (OS) (6-8). However, not all patients respond well to NACT (6). Considering the side-effect of chemotherapy and the requirement for timely and effective treatment for early breast cancer patients, early prediction of NACT responders has emerged as a critical clinical issue.

Breast magnetic resonance imaging (MRI) was first introduced to clinical practice during the early 1990s. It shows excellent diagnostic accuracy compared to mammography or ultrasound and is considered the most accurate detection means for breast diseases (9). Recently, emerging evidence highlights the prognostic utility of multiparametric MRI in predicting NACT response. Several parameters of MRI, such as morphology (10), volume transfer constant (Ktrans), and early contrast uptake (ECU) (11), have been shown to be predictable for pCR.

Conventional assessment of NACT response has primarily relied on tumor dimensional measurements (12), but it has two fundamental limitations: (I) limited sensitivity for early treatment response detection; and (II) inapplicability to cytostatic agents that modulate tumor proliferation without killing tumor cells immediately (13). Functional imaging technologies, which evaluate tumor angiogenesis, metabolic, and molecule changes are more sensitive (14). Diffusion-weighted imaging (DWI), employing pulsed gradient sequences to quantify Brownian motion of water molecules, expressed as apparent diffusion coefficient (ADC) values, provides unique insights into the integrity of cell membranes and cellularity. In addition to being noninvasive, with no radiation exposure, and providing both morphologic and physiologic information, it does not require a contrast medium, which is indispensable in contrast-enhanced MRI (15). ADC is considered a potential biomarker for NACT response, since cell lysis caused by chemotherapy leads to less restriction for water diffusion and therefore a higher ADC (16,17).

A recent meta-analysis demonstrated no significant predictive value of pretreatment ADC for NACT response, with significant overlapping between responders [0.98; 95% confidence interval (CI): 0.94–1.03] and non-responders (1.05; 95% CI: 1.00–1.10) (18). However, whether ADC change after 1 or 2 cycles of NACT has the prediction ability is unknown. ADC change was previously reported to identify responders effectively (16,19), and this change occurs earlier than tumor size (20), indicating that it may be an effective parameter to predict response to NACT in the early stage. Percentage ADC change {ΔADC%=[(ADCpost−ADCpre)/ADCpre]×100%}, can reduce inter-scanner variability, and several studies have shown its role in this field (21-23). Nevertheless, the relatively small size of samples restricts the statistical power of current studies. In this study, we aim to conduct a meta-analysis and a retrospective validation study to evaluate the role of early ΔADC% in predicting response to NACT in patients with breast cancer, potentially guiding NACT strategies in the early stage. We present this article in accordance with the PRISMA reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1167/rc).

Methods

Meta-analysis

Literature search and selection

PubMed, Web of Science, and Cochrane Library were searched for original studies for this meta-analysis. The search was updated until April 28, 2025. Search terms included “breast cancer”, “breast tumor”, “breast carcinoma”, “breast neoplasm”, “mammary cancer”, “DWI”, “diffusion weighted imaging”, “multiparametric MRI”, “multiparametric magnetic resonance imaging”, “ADC”, “apparent diffusion coefficient”, “diffusion weighted MR”, and “neoadjuvant chemotherapy”. The references of the included studies were also searched.

The inclusion criteria were: predicting response to NACT in breast cancer; using ΔADC% as prediction parameter; DWI-MRI images were acquired before NACT beginning and early during NACT (after 1 or 2 cycles); providing exact criteria for responders.

The exclusion criteria were: not in English or Chinese; animal studies; review/meta-analysis/abstract/case report; the number of true-positive (TP), false-negative (FN), false-positive (FP), and true-negative (TN) findings cannot be extracted directly or indirectly.

Study selection

Duplicated studies were removed after literature search. Two independent authors screened out the remaining studies by reading the titles and abstracts. Full-text review was performed independently by the same two authors according to the inclusion and exclusion criteria. Disagreements were resolved through discussion. If the divergence could not be resolved, a third review author was consulted.

Data extraction

A data extraction spreadsheet was developed. Two independent authors extracted data according to the spreadsheet. We extracted basic study characteristics, including publication year, number of patients, age of patients (mean/median age and age range), criteria of responders, and study design. We also extracted MRI and diagnostic parameters, including B₀-field strength, b values, time of MRI scans, cut-off value, sensitivity, specificity, and the number of TP, FN, FP, and TN.

For studies that reported ΔADC% both after 1 and 2 cycles of NACT, we extracted the data separately. For studies having overlapping patients, we only extracted data from one study.

Data quality assessment

The methodological quality of each included study was assessed by the Quality Assessment of Diagnostic Accuracy Studies (QUADAS)-2 tool (24) in Revman (version 5.4.1; The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark). Degree of heterogeneity was evaluated by Cochran Q test and Higgins I² test calculating by Stata (version 16.0; StataCorp, College Station, TX, USA). Publication bias was assessed by a Deeks’ funnel plot using Stata.

Statistical analysis

We constructed forest plots for sensitivity and specificity for each individual study included. To show the pooled effect of early ΔADC% in predicting response to NACT, summary receiver operating characteristic (SROC) curve was constructed and the area under the curve (AUC) was calculated. Bivariate boxplot was used to evaluate the heterogeneity and threshold effects. To find the possible heterogeneity source, we performed subgroup analyses. All data were provided with 95% CI. All these analyses were conducted in Stata 16.0.

Validation study

Patients

A total of 47 patients with breast-invasive carcinomas from Wuhan Union Hospital were retrospectively studied. The inclusion criteria were: female patients; biopsy-proven breast-invasive carcinomas; underwent standard NACT; DWI-MRI performed before and after 1 or 2 cycles of NACT; and mastectomy or breast-conserving surgery was conducted after NACT. Clinical and histological data, including age, menstrual status, and molecular subtype, were collected. The patients who reached pCR after NACT were labeled as the pCR group, and those who did not were labeled as the non-pCR group. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology (No. 0248) and the requirement for written informed consent was waived by institutional policy and the retrospective study design.

MRI data acquisition

MR images were acquired using SIEMENS Verio 3.0 T MR imaging system with a 16-channel breast surface coil. During examination, patients were in a prone position. DWI parameters are as following: fat suppressed, spectral attenuated inversion-recovery (SPAIR); repetition time (TR), 4,760 ms; echo time (TE), 57 ms; voxel size, 1.5 mm × 1.5 mm × 3.0 mm; field of view (FOV), 192×320 mm²; flip angle, 180°; b values, 0 and 1,000 s/mm²; acquisition time, 209 s; number of images, 42; number of excitation, 1 at a b-value of 50 s/mm² and 2 at a b-value of 1,000 s/mm².

Image analysis

A single region of interest (ROI, 30–50 mm²) was manually drawn inside the tumor on the ADC map with the largest cross-section of the lesion on the b=0 and 1,000 s/mm² images. ADC was recorded for each patient.

Statistical analysis

Statistical analysis of the retrospective study was performed by R (version 4.1.2, R Foundation for Statistical Computing, Vienna, Austria). Continuous variables were presented as mean ± standard deviation (SD), and categorical variables were presented as number and percentage. ΔADC% is defined as follows:

ΔADC%=(ADC2−ADC1)/ADC1×100%[1]

ADC1: ADC of pretreatment DWI.

ADC2: ADC of early treatment (after 1 or 2 cycles of NACT) DWI.

We performed Mann-Whitney U test to compare the ΔADC% and ADC1 between the pCR group and the non-pCR group. P values of less than 0.05 were considered statistically significant. Receiver operating characteristic (ROC) curve analysis and its best cut-off point were used to show the prediction ability of ΔADC%.

Results

Meta-analysis

Study selection and data extraction

The study selection process was schematically outlined in Figure 1, which illustrated the systematic screening protocol employed in this meta-analysis. Search strategies incorporating keywords and Boolean operators were comprehensively detailed in the Methods section. The search initially yielded 590 potentially relevant records through the predefined search algorithm from PubMed, Web of Science, and Cochrane Library after removing duplicated studies. Following rigorous title and abstract screening, 529 studies were excluded. By full-text screening of the remaining 61 studies, 11 studies (21-23,25-32) meeting all eligibility criteria for quantitative synthesis were finally included. There were two main reasons for exclusion: 28 studies were eliminated due to insufficient diagnostic data (absence of TP/FP/FN/TN values), while 22 studies were excluded because of methodological incompatibility, which used other parameters other than ΔADC%. Table 1 provided comprehensive details of the included studies, encompassing demographic characteristics, methodological details, diagnostic contingency data, imaging acquisition parameters, and diagnostic threshold specifications. It is notable that Xu et al. (27) analyzed the data both after 1 and 2 cycles of NACT, and in this study, we analyzed them separately.

Figure 1 Flowchart for selection and exclusion of studies.

Data quality assessment

The methodological quality assessment of included studies, as evaluated through the QUADAS-2 tool, was systematically presented in Figure 2. There were no applicability concerns. The risk of bias with respect to patient selection was rated unclear in six studies (25-28,30,32) due to the absence of explicit statements regarding consecutive or random enrollment. The risk of bias with respect to the index test was rated unclear in five studies (22,25,26,28,30) because they did not report whether index test interpretation was blinded to reference standard results. The risk of bias with respect to reference standard was rated unclear in six studies (22,23,25,27,28,30) for two reasons: (I) lack of reporting on blinding between reference standard and index test interpretations (22,23,25,27,28,30); and (II) utilization of non-validated diagnostic criteria (pathological response omitted as reference standard (22,27,28). The risk of bias with respect to flow and timing was rated high in 1 study because not all patients were included in the study (30).

Figure 2 Study quality evaluated by QUADAS-2. QUADAS, Quality Assessment of Diagnostic Accuracy Studies.

The evaluation of publication bias through Deeks’ funnel plot asymmetry test demonstrated no statistically significant asymmetry (P=0.843), as visually evidenced by the symmetrical distribution of study effect estimates in Figure 3.

Figure 3 Publication bias assessed by Deeks’ funnel plot. ESS, effective sample size.

Data analysis

The diagnostic performance of ΔADC% in predicting the response to NACT was visually summarized through the bivariate forest plot presented in Figure 4. Quantitative heterogeneity analysis revealed substantial between-study variance. Sensitivity estimates demonstrated significant heterogeneity (I²=84.02%; 95% CI: 75.98–92.05; Cochran’s Q=68.82; P<0.005). Specificity estimates exhibited moderate-to-high heterogeneity (I²=72.20%; 95% CI: 56.04–88.37; Cochran’s Q=39.57, P<0.005). Figure 5 illustrated the bivariate boxplot of sensitivity and specificity across included studies. The outlier studies illustrated the existence of heterogeneity, which was consistent with the I² and Cochran’s Q data. Besides, there was no significant positive or negative correlation between sensitivity and specificity, suggesting a low likelihood of threshold effects.

Figure 4 Forest plots of sensitivity and specificity with 95% CIs per study. CI, confidence interval.

Figure 5 Bivariate boxplot of sensitivity and specificity. SENS, sensitivity; SPEC, specificity.

The SROC curve presented in Figure 6 demonstrated the diagnostic accuracy profile of the pooled analysis. The bivariate random-effects model yielded a summary sensitivity of 0.71 (95% CI: 0.60–0.79) and specificity of 0.78 (95% CI: 0.68–0.85), with AUC reaching 0.81 (95% CI: 0.77–0.84), indicating good discriminative capacity.

Figure 6 SROC curve with 95% CI regions. AUC, area area under the curve; CI, confidence interval; SENS, sensitivity; SPEC, specificity; SROC, summary receiver operating characteristic.

Given the significant heterogeneity observed across studies (I²=82%; P<0.01), multivariate meta-regression and subgroup analyses were systematically conducted to explore potential moderators. As shown in Table 2 and Figure 7, no statistically significant differences were identified among key methodological and technical covariates (P>0.05), indicating these parameters did not contribute to the heterogeneity.

Figure 7 Univariable meta-regression and subgroup analyses for study design, magnetic field intensity, maximum b value, time of the second MRI scan, and criteria for responders. CI, confidence interval; MRI, magnetic resonance imaging; pCR, pathological complete response.

Validation study

Patient characteristics

The study cohort comprised 47 patients with a median age of 49 (range, 28–68) years. In total, 22 (46.8%) patients were premenopausal, 21 (44.7%) were postmenopausal, and menopausal status of 4 (8.5%) patients was unknown. Pathological evaluation identified 17 patients (36.2%) achieving pCR, while 30 patients (63.8%) were classified as non-responders. As for molecular subtype, 4 (8.5%) cancers were triple-negative [estrogen receptor (ER)-negative, human epidermal growth factor receptor 2 (HER2)-negative, and progesterone receptor (PR)-negative], 20 (42.6%) were luminal (ER-positive), and 23 (48.9%) were HER2 enriched (ER-negative and HER2-positive). A comprehensive summary of clinicopathological characteristics was presented in Table S1. Representative MRI images of pre-treatment and after 1–2 cycles of NACT were shown in Figure 8.

Figure 8 Representative MRI images of pre-treatment and after 1–2 cycles of NACT. MRI, magnetic resonance imaging; NACT, neoadjuvant chemotherapy; pCR, pathological complete response.

Prediction of treatment responses

Quantitative analysis demonstrated significantly elevated ΔADC% in the pCR group compared to non-responders (42.0%±40.2% vs. 17.6%±21.7%, P=0.032). In contrast, baseline pre-treatment ADC values showed no statistically significant intergroup difference (0.79×10⁻³±0.11×10⁻³ vs. 0.78×10⁻³±0.11×10⁻³ mm²/s, P=0.833). Furthermore, ROC revealed distinct diagnostic performance, and the AUC of baseline ADC was 0.520, indicating limited predictive capacity, while the AUC of ΔADC% was 0.690, demonstrating moderate discriminative power. A cut-off value of 20.3% for ΔADC% achieved a sensitivity of 64.7% and a specificity of 73.3% for pCR prediction (Figure 9A). The sensitivity and specificity corresponding to different ADCs were detailed in Table S2. It is interesting to find that there was a weak correlation between ΔADC% and baseline ADC (Figure 9B). However, since the sample was relatively small in our study, their relationship needs validation in larger cohorts.

Figure 9 Comparation of the diagnostic value of ΔADC% and baseline ADC. (A) ROC curves of ΔADC% and baseline ADC of the validation cohort. (B) Linear correlation analysis of ΔADC% and baseline ADC. ADC1: ADC of pretreatment DWI. ΔADC%, percentage apparent diffusion coefficient change; ADC, apparent diffusion coefficient; DWI, diffusion-weighted imaging; ROC, receiver operating characteristic.

Discussion

The study systematically evaluated the predictive performance of ΔADC% for NACT response in breast cancer. The meta-analysis demonstrated robust diagnostic accuracy with the pooled sensitivity and specificity are 0.71 (95% CI: 0.60–0.79) and 0.78 (95% CI: 0.68–0.85), respectively, and the AUC is 0.81 (95% CI: 0.77–0.84), indicating ΔADC% as a valuable image marker to predict NACT response after only 1 or 2 treatment cycles. The retrospective validation study verified the results. ΔADC% in the pCR group is significantly higher than the non-pCR group, and when the ADC cut-off value was 20.3%, the sensitivity and specificity for predicting pCR are 64.7% and 73.3%, respectively. The results may provide new methods to identify responders in an early stage and are helpful in guiding personalized therapeutic strategies.

Several other meta-analyses have assessed the role of DWI in predicting NACT response in breast cancer. Gao et al. (33) evaluated DWI in diagnosing pCR for patients with breast cancer who underwent NACT treatment and reported a sensitivity of 0.89 (95% CI: 0.86–0.91) and a specificity of 0.72 (95% CI: 0.68–0.75). Similarly, Li et al. (34) performed a head-to-head comparison of DWI vs. DCE-MRI across 41 studies, demonstrating superior diagnostic performance of DWI. However, both reviews employed broad temporal classifications. Although subgroup analyses were conducted, there are only three subgroups, namely pre-NACT, mid-NACT, and post-NACT, without separate analysis of early treatment in NACT (after 1 or 2 cycles). Notably, Surov et al. (18) focused on pretreatment ADC specifically but found that it has limited predictive capacity since the ADC values of the responders and non-responders overlapped significantly. Our validation study accords with their results that ADC of pretreatment MRI cannot distinguish pCR patients from non-pCR patients. Possibly, it is because ADC reflects cellularity and interstitial water mobility, which is changed by NACT-caused cell lysis.

To our knowledge, this is the first meta-analysis to investigate whether early ΔADC% can predict response to NACT, and our results show its value in this field. Compared to those studies focused on post-NACT, our results help to find the responders earlier. Meanwhile, compared to those focused on pre-NACT, our results show better prediction ability.

Despite demonstrating promising diagnostic potential, significant heterogeneity was noticed in our study. Sub-group analyses were conducted but did not show obvious heterogeneity in study design (prospective or retrospective), magnetic field strength (1.5 or 3.0 T), criteria of responders (clinical or pathological), maximum b-value (≥1,000 or <1,000 s/mm²), and time of the second scan (after 1 or 2 cycles of NACT). Some other factors, such as tumor size, subtype, and chemotherapy regimen, may explain the heterogeneity. Regrettably, insufficient reporting of these variables prevents us from further investigation. This methodological limitation underscores the need for prospective trials incorporating standardized protocols to further illustrate the source of heterogeneity. Our study has some other limitations. First, the statistical power of both the meta-analysis and validation cohort is constrained by limited sample sizes, ranging from 24 to 174. Thus, a larger population is required to confirm the results. Second, all the studies include patients with different molecular subtypes of breast cancer. Previous studies showed that there is a significant ADC difference among breast cancer subtypes according to distinct tumor microenvironment characteristics - including variations in angiogenesis, necrosis patterns, and cellular density (35). The prediction accuracy of ADC to predict pCR is also associated with tumor subtypes, and the performance is better in the triple-negative and HER2-positive subgroups (36). Therefore, subtype-optimized ADC thresholds are required. Moreover, ADC itself has inherent limitations. The ADC value measured in vivo is more influenced by T2 effects and does not purely represent actual tissue diffusion (37). Thus, more accurate imaging methodologies are necessary for predicting pCR.

Conclusions

This study demonstrates that early ΔADC% exhibits potential as a quantitative imaging biomarker for predicting NACT response in breast cancer, with implications for guiding personalized therapeutic decision-making. However, considering the small size and obvious heterogeneity of the included studies, prospective validation through multicenter trials incorporating standardized acquisition protocols and subtype-stratified analyses is required to establish clinical utility.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the PRISMA reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1167/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1167/dss

Funding: This work was supported by the National Natural Science Foundation of China (No. 82270830 to J.M.), Nature Science Foundation of Hubei Province (No. 2022CFB072 to J.M. and No. 2022CFB230 to J.L.), and Hubei Province Key Laboratory of Molecular Imaging (No. 2024FZYX022 to J.M.).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1167/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology (No. 0248) and the requirement for written informed consent was waived by institutional policy and the retrospective study design.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, Jemal A. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Mougalian SS, Soulos PR, Killelea BK, Lannin DR, Abu-Khalaf MM, DiGiovanna MP, Sanft TB, Pusztai L, Gross CP, Chagpar AB. Use of neoadjuvant chemotherapy for patients with stage I to III breast cancer in the United States. Cancer 2015;121:2544-52. [Crossref] [PubMed]
3. Clough KB, Acosta-Marín V, Nos C, Alran S, Rouanet P, Garbay JR, Giard S, Verhaeghe JL, Houvenaeghel G, Flipo B, Dauplat J, Dorangeon PH, Classe JM, Rouzier R, Bonnier P. Rates of Neoadjuvant Chemotherapy and Oncoplastic Surgery for Breast Cancer Surgery: A French National Survey. Ann Surg Oncol 2015;22:3504-11. [Crossref] [PubMed]
4. Hart SE, Brown DL, Kim HM, Qi J, Hamill JB, Wilkins EG. Association of Clinical Complications of Chemotherapy and Patient-Reported Outcomes After Immediate Breast Reconstruction. JAMA Surg 2021;156:847-55. [Crossref] [PubMed]
5. Long-term outcomes for neoadjuvant versus adjuvant chemotherapy in early breast cancer: meta-analysis of individual patient data from ten randomised trials. Lancet Oncol 2018;19:27-39. [Crossref] [PubMed]
6. Cortazar P, Geyer CE Jr. Pathological complete response in neoadjuvant treatment of breast cancer. Ann Surg Oncol 2015;22:1441-6. [Crossref] [PubMed]
7. von Minckwitz G, Untch M, Blohmer JU, Costa SD, Eidtmann H, Fasching PA, Gerber B, Eiermann W, Hilfrich J, Huober J, Jackisch C, Kaufmann M, Konecny GE, Denkert C, Nekljudova V, Mehta K, Loibl S. Definition and impact of pathologic complete response on prognosis after neoadjuvant chemotherapy in various intrinsic breast cancer subtypes. J Clin Oncol 2012;30:1796-804. [Crossref] [PubMed]
8. Cortazar P, Zhang L, Untch M, Mehta K, Costantino JP, Wolmark N, et al. Pathological complete response and long-term clinical benefit in breast cancer: the CTNeoBC pooled analysis. Lancet 2014;384:164-72. [Crossref] [PubMed]
9. Kuhl CK. Abbreviated Magnetic Resonance Imaging (MRI) for Breast Cancer Screening: Rationale, Concept, and Transfer to Clinical Practice. Annu Rev Med 2019;70:501-19. [Crossref] [PubMed]
10. Michishita S, Kim SJ, Shimazu K, Sota Y, Naoi Y, Maruyama N, Kagara N, Shimoda M, Shimomura A, Noguchi S. Prediction of pathological complete response to neoadjuvant chemotherapy by magnetic resonance imaging in breast cancer patients. Breast 2015;24:159-65. [Crossref] [PubMed]
11. Marinovich ML, Sardanelli F, Ciatto S, Mamounas E, Brennan M, Macaskill P, Irwig L, von Minckwitz G, Houssami N. Early prediction of pathologic response to neoadjuvant therapy in breast cancer: systematic review of the accuracy of MRI. Breast 2012;21:669-77. [Crossref] [PubMed]
12. Litière S, Collette S, de Vries EG, Seymour L, Bogaerts J. RECIST - learning from the past to build the future. Nat Rev Clin Oncol 2017;14:187-92. [Crossref] [PubMed]
13. Tuma RS. Sometimes size doesn't matter: reevaluating RECIST and tumor response rate endpoints. J Natl Cancer Inst 2006;98:1272-4. [Crossref] [PubMed]
14. Rauch GM, Adrada BE, Kuerer HM, van la Parra RF, Leung JW, Yang WT. Multimodality Imaging for Evaluating Response to Neoadjuvant Chemotherapy in Breast Cancer. AJR Am J Roentgenol 2017;208:290-9. [Crossref] [PubMed]
15. Abdel Razek AA, Gaballa G, Denewer A, Tawakol I. Diffusion weighted MR imaging of the breast. Acad Radiol 2010;17:382-6. [Crossref] [PubMed]
16. Partridge SC, Zhang Z, Newitt DC, Gibbs JE, Chenevert TL, Rosen MA, Bolan PJ, Marques HS, Romanoff J, Cimino L, Joe BN, Umphrey HR, Ojeda-Fournier H, Dogan B, Oh K, Abe H, Drukteinis JS, Esserman LJ, Hylton NM. Diffusion-weighted MRI Findings Predict Pathologic Response in Neoadjuvant Treatment of Breast Cancer: The ACRIN 6698 Multicenter Trial. Radiology 2018;289:618-27. [Crossref] [PubMed]
17. Lehman CD. Diffusion weighted imaging (DWI) of the breast: ready for clinical practice? Eur J Radiol 2012;81:S80-1. [Crossref] [PubMed]
18. Surov A, Wienke A, Meyer HJ. Pretreatment apparent diffusion coefficient does not predict therapy response to neoadjuvant chemotherapy in breast cancer. Breast 2020;53:59-67.
19. Bedair R, Priest AN, Patterson AJ, McLean MA, Graves MJ, Manavaki R, Gill AB, Abeyakoon O, Griffiths JR, Gilbert FJ. Assessment of early treatment response to neoadjuvant chemotherapy in breast cancer using non-mono-exponential diffusion models: a feasibility study comparing the baseline and mid-treatment MRI examinations. Eur Radiol 2017;27:2726-36. [Crossref] [PubMed]
20. Pickles MD, Gibbs P, Lowry M, Turnbull LW. Diffusion changes precede size reduction in neoadjuvant treatment of breast cancer. Magn Reson Imaging 2006;24:843-7. [Crossref] [PubMed]
21. Pereira NP, Curi C, Osório CABT, Marques EF, Makdissi FB, Pinker K, Bitencourt AGV. Diffusion-Weighted Magnetic Resonance Imaging of Patients with Breast Cancer Following Neoadjuvant Chemotherapy Provides Early Prediction of Pathological Response - A Prospective Study. Sci Rep 2019;9:16372. [Crossref] [PubMed]
22. Wu LA, Chang RF, Huang CS, Lu YS, Chen HH, Chen JY, Chang YC. Evaluation of the treatment response to neoadjuvant chemotherapy in locally advanced breast cancer using combined magnetic resonance vascular maps and apparent diffusion coefficient. J Magn Reson Imaging 2015;42:1407-20. [Crossref] [PubMed]
23. Li X, Abramson RG, Arlinghaus LR, Kang H, Chakravarthy AB, Abramson VG, Farley J, Mayer IA, Kelley MC, Meszoely IM, Means-Powell J, Grau AM, Sanders M, Yankeelov TE. Multiparametric magnetic resonance imaging for predicting pathological response after the first cycle of neoadjuvant chemotherapy in breast cancer. Invest Radiol 2015;50:195-204. [Crossref] [PubMed]
24. Whiting PF, Rutjes AW, Westwood ME, Mallett S, Deeks JJ, Reitsma JB, Leeflang MM, Sterne JA, Bossuyt PM. QUADAS-2 Group. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011;155:529-36. [Crossref] [PubMed]
25. Sharma U, Agarwal K, Sah RG, Parshad R, Seenu V, Mathur S, Gupta SD, Jagannathan NR. Can Multi-Parametric MR Based Approach Improve the Predictive Value of Pathological and Clinical Therapeutic Response in Breast Cancer Patients? Front Oncol 2018;8:319. [Crossref] [PubMed]
26. Choi JH, Kim HA, Kim W, Lim I, Lee I, Byun BH, Noh WC, Seong MK, Lee SS, Kim BI, Choi CW, Lim SM, Woo SK. Early prediction of neoadjuvant chemotherapy response for advanced breast cancer using PET/MRI image deep learning. Sci Rep 2020;10:21149. [Crossref] [PubMed]
27. Xu HD, Zhang YQ. Evaluation of the efficacy of neoadjuvant chemotherapy for breast cancer using diffusion-weighted imaging and dynamic contrast-enhanced magnetic resonance imaging. Neoplasma 2017;64:430-6. [Crossref] [PubMed]
28. Sharma U, Danishad KK, Seenu V, Jagannathan NR. Longitudinal study of the assessment by MRI and diffusion-weighted imaging of tumor response in patients with locally advanced breast cancer undergoing neoadjuvant chemotherapy. NMR Biomed 2009;22:104-13. [Crossref] [PubMed]
29. Hottat NA, Badr DA, Lecomte S, Besse-Hammer T, Jani JC, Cannie MM. Value of diffusion-weighted MRI in predicting early response to neoadjuvant chemotherapy of breast cancer: comparison between ROI-ADC and whole-lesion-ADC measurements. Eur Radiol 2022;32:4067-78. [Crossref] [PubMed]
30. Cao MR, Liu BG, Huang Y, et al. The Inchoate Predictive Value of Diffusion Weighted MR Imaging in Neoadjuvant Chemotherapy Effect of Breast Cancer. Journal of Clinical Radiology 2012;31:1554-8.
31. Du S, Gao S, Zhao R, Liu H, Wang Y, Qi X, Li S, Cao J, Zhang L. Contrast-free MRI quantitative parameters for early prediction of pathological response to neoadjuvant chemotherapy in breast cancer. Eur Radiol 2022;32:5759-72. [Crossref] [PubMed]
32. He M, Su J, Ruan H, Song Y, Ma M, Xue F. Nomogram based on quantitative dynamic contrast-enhanced magnetic resonance imaging, apparent diffusion coefficient, and clinicopathological features for early prediction of pathologic complete response in breast cancer patients receiving neoadjuvant chemotherapy. Quant Imaging Med Surg 2023;13:4089-102. [Crossref] [PubMed]
33. Gao W, Guo N, Dong T. Diffusion-weighted imaging in monitoring the pathological response to neoadjuvant chemotherapy in patients with breast cancer: a meta-analysis. World J Surg Oncol 2018;16:145. [Crossref] [PubMed]
34. Li Z, Li J, Lu X, Qu M, Tian J, Lei J. The diagnostic performance of diffusion-weighted imaging and dynamic contrast-enhanced magnetic resonance imaging in evaluating the pathological response of breast cancer to neoadjuvant chemotherapy: A meta-analysis. Eur J Radiol 2021;143:109931. [Crossref] [PubMed]
35. Liu S, Ren R, Chen Z, Wang Y, Fan T, Li C, Zhang P. Diffusion-weighted imaging in assessing pathological response of tumor in breast cancer subtype to neoadjuvant chemotherapy. J Magn Reson Imaging 2015;42:779-87. [Crossref] [PubMed]
36. Bufi E, Belli P, Di Matteo M, Terribile D, Franceschini G, Nardone L, Petrone G, Bonomo L. Effect of breast cancer phenotype on diagnostic performance of MRI in the prediction to response to neoadjuvant treatment. Eur J Radiol 2014;83:1631-8. [Crossref] [PubMed]
37. Wáng YXJ. An explanation for the triphasic dependency of apparent diffusion coefficient (ADC) on T2 relaxation time: the multiple T2 compartments model. Quant Imaging Med Surg 2025;15:3779-91. [Crossref] [PubMed]