4.6.2. Breast

The DWI and magnetic resonance spectroscopy (MRS) were able to detect early tumor responses in triple-negative breast cancer (TNBC) mouse xenografts after combination therapy with TRA-8, a monoclonal antibody targeted to the apoptosis receptor, and carboplatin, a standard chemotherapeutic agent. The therapeutic efficacy was assessed by monitoring tumor volume, first ADC changes, and lipid concentration through the fat–water ratio (FWR) MRS. Two different TNBC xenograft models, implanted either with 2LMP or SUM159 cell lines, were treated intraperitoneally with either carboplatin, TRA-8, or their combination. The MRI acquisitions with DWI and MRS were performed before, during, and at the end of the therapeutic protocol. Combination therapy with TRA-8 and carboplatin showed significantly reduced tumor growth in both 2LMP and SUM159 TNBC models compared to both single-drug treatments, and the therapeutic efficacy was verified histologically, as witnessed by a significant increase of apoptotic cell density. Substantial changes in ADC were detected only three days after the initiation of TRA-8 or combination therapy, while significant changes in FWR required seven days for detection. Both ADC and FWR changes were confirmed as useful imaging biomarkers to evaluate the therapeutic efficacy, but ADC changes may be used as an earlier predictor. This imaging protocol might be translated into clinical trials testing these or similar drugs, since early assessment is useful for preventing unnecessary/ineffective treatments and, hence, to improve outcomes [146].

In the preclinical model of ductal breast cancer with BT474 cells previously described (see 4.1 T1 and T2 Weighted Sequences, [109]), ADC was also tested as a predictor of gefitinib efficacy. Tumor ADC increased in all treated groups. Therefore, as mentioned, the parameters included in the study, specifically *Kps* and *fPV* from DCE and ADC from DWI, may all provide, in combination, a more sensitive approach to determining the right treatment plan [109].

A mouse model of a breast cancer brain metastasis was developed to optimize a longitudinal MRI and MRS method for analyzing the pathogenic aspects of brain metastasis, since cerebral metastasis showed a high incidence in breast carcinoma patients. Brain metastases were generated by the intra-carotid injection of human mammary carcinoma cells 435-Br1, a brain metastasis variant of the parental cell line MDA-MB-435. Different MRI approaches were used to characterize the morphological and metabolic development patterns of brain metastases, including T2 and contrast enhanced-T1 weighted images, DWI, and MRS imaging. The acquired data were then combined with the histological analyses of the dissected brains. As a result, nine out of thirteen mice developed MR-detectable abnormal masses (single or multiple) in the brain parenchyma within 20 to 62 days after injection. Two additional animals had brain metastases detected only by histological analysis. The ADC maps correctly differentiated the edema and disorders of cerebrospinal fluid circulation areas from metastases, in addition to the MRS results that have been discussed in the dedicated paragraph. This imaging approach might be considered helpful in early detection and differentiation of brain metastases from other cerebral abnormalities [147].

In the preliminary study of breast cancer in human patients already described (see 4.2 Dynamic Contrast-Enhanced [122]), the DWI method was used to predict tumors' responses following neoadjuvant therapy (NAT), in conjunction with DCE. After completion of NAT, the pathologic complete response (pCR) and nonresponse (non-pCR) were determined at the time of surgery. The mean ADC was increased between the first two time points for the same pCR, while no noticeable change in mean ADC was observed for the non-pCR patient. The MRI approach described may improve the accuracy of evaluating the tumor response to NAT, showing a higher predictive power than models based on tumor size changes, and it may be used as an early marker of outcome in breast cancer patients [122].

## 4.6.3. Prostate

In a prostate cancer xenograft mice model, DWI was tested as a predictor of the efficacy of docetaxel, a cytostatic and cytotoxic antineoplastic drug. The LnCaP cell line, derived from the lymph node metastasis of a human prostatic adenocarcinoma was used, since it secretes the prostate-specific antigen (PSA). Serum PSA levels and tumor volumes were used to measure tumor response noninvasively. Results showed a good response of the xenograft to docetaxel, and serum PSA was confirmed as a useful biomarker of therapy response, but changes in ADC represented an earlier indicator of therapeutic response. This approach should be validated in the clinic to correctly manage the therapeutic regimen, especially for second-line therapies that show reduced response rates and higher toxicity [148].

In the study described before (see 4.2 Dynamic Contrast-Enhanced [123]), mice bearing prostate cancer xenografts underwent, as mentioned, DWI acquisitions pre-radiation, and on day one and nine after. The ADC values, volumes, and PSA levels revealed significant correlations with treatment response. The combination of all parameters, including the *Ktrans* obtained with the DCE, successfully predicted treatment response with a high correlation coefficient. Therefore, the combination of such parameters with standard clinical parameters may improve the predictive power for the therapeutic outcome in prostate cancer patients, and it might be applied for the personalization of therapeutic approaches [123].

A multiparametric analysis to evaluate in vivo vascularization, metabolism, and physiological characteristics that are permissive for the occurrence of metastases has been tested in a prostate cancer xenograft models using noninvasive MRI and MRS. A human prostate cancer xenograft model implanted s.c. or orthotopically in the prostate was used. The tumors were derived from PC-3 human prostate cells inoculated s.c. in the right flank of SCID mice; the intact tumor tissue obtained from subcutaneous tumors was harvested and re-implanted s.c. in the flank or orthotopically with a microsurgical method of SCID mice. These mouse models underwent MRI acquisitions, including T1-weighted with the administration of albumin-Gd-DTPA, multislice DWI for evaluation of vasculature, and MRS for metabolic maps of total choline and lactate/lipid and extracellular pH (pHe). As the results showed, the tumors in the orthotopic site were identifiable by the hyperintense signal detected in DWI compared to the s.c., and they showed higher vascular volume and higher permeability, as well. In the metastases' characterizations, several lung nodules were observed in mice with orthotopically-implanted tumors, and only a few small clusters of cells were observed in the lungs of mice with xenografts. Therefore, the higher metastatization rate of orthotopic tumors might be linked

to the higher vascular volume and permeability. These data confirm the profound influence of the tumor microenvironment on the metastatization process and the feasibility of identifying noninvasive clinically translatable parameters that may contribute to determining risk factors for metastases formation in patients. Future translational studies are necessary to validate these observations further [149].

#### *4.7. Di*ff*usion Kurtosis Imaging*

Diffusion kurtosis imaging (DKI) has been applied to many body-districts, showing a higher capability in differentiating between benign and malignant tumors, and in predicting treatment responses compared to conventional DWI. From the DKI model, the diffusion coefficient (D) and the diffusional kurtosis (K) are derived, which provide more information on the tissue structure than DWI. Indeed, the parameter D evaluates the tissue diffusion and seems to be more accurate than ADC; the K parameter seems to be more sensitive to irregular and heterogeneous cellular environments, but less sensitive to cell density.
