br Radiomics there will be a continued search for
Radiomics - there will be a continued search for quantitative image features having observable biological correlates
Radiomics is a recent buzzword for the extraction and applica-tion of a large number of quantitative image features. Radiomics can potentially be used to improve diagnosis, characterization and prognostication of breast cancer . It has mainly been employed in the more information-rich modality of MRI, and to certain extent PET imaging. Radiomics was inspired by the other omics, e.g., ge-nomics, proteomics, but there is an important difference. Many of the features in the other omics stem from identifiable biological entities, such as the DNA or the proteins in the body. The radiomics features are however more abstract statistical and mathematical features of images of the body, tissues and cells. We believe that to differentiate itself from AI and deep learning, radiomics research will focus on identifying image features robustly associated with observable biological entities . Minimally invasive image-guided procedures will replace open surgery for smaller lesions
Early detection of breast cancer means a shift towards smaller tumor sizes at diagnosis which put a demand on minimal surgical intervention to keep surgery-related morbidity as low as possible. Alongside, technical development has produced several minimally invasive imaging and interventional candidate techniques such as ablation based on cryotherapy, laser, microwaves, high intensity focused ultrasound (HIFU) and radiofrequency. Most of them require a minimally invasive needle L-Glutamine although HIFU may be performed non-invasively. A minimally invasive technique can move the surgical procedure from the operating room to an “office setting”, be performed under local anesthesia and thus minimize the need for perioperative resources. A correct tumor visualization and delineation (roadmap) and image guidance during the pro-cedure is crucial. In addition to traditional imaging such as mammography and ultrasound, MRI may be required to select the appropriate patient group and ultrasound (perhaps contrast-enhanced) may be the best way to monitor the invasive proced-ure in real time. Minimally invasive intervention is of interest both for benign and high-risk lesions, but also for small, malignant le-sions foremost in elderly patients or patients with comorbidities where traditional surgery is not an option. The challenge of accepting the fact that the evidence (the tumor) is vanished and not available for pathological investigation with the majority of these techniques after treatment poses high demands on selecting the proper patient group for these procedures. Still, minimal invasive interventions are of great interest as we move towards detection of smaller tumors with more sensitive diagnostic techniques.
Precise guidelines for MRI in preoperative mapping will be established
The goal of breast cancer surgery is to remove the malignant tissue in a tailored manner. The proportion of women receiving breast-conserving surgery is high. When there are positive or minimal margins the patient must suffer a second surgery. In a recent randomized study including all recently diagnosed women below 56 years of age, Immortalization was shown that MRI changed the treat-ment plan in 18% of the patients whilst the re-excision rate after breast conserving surgery was 22% without MRI and 5% with MRI . The use of MRI leads to an increased number of biopsies. One study found that around 50% of additional findings on MRI could subsequently be located by second look ultra-sound and an addi-tional 25% by tomosynthesis. Of the remaining additional findings which would require an MRI biopsy procedure to confirm, 4 out of 5 corresponded to malignant lesions. Long-term follow-up of 470 patients showed that the in-breast recurrence rate was 4.2% without MRI and 1.6% with MRI . At the same time, an individual-level meta-analysis showed no difference in local recurrence comparing MRI with no MRI . We are still waiting for final results on the impact on mortality . When examining radiotherapy, it has been shown that a 5-year difference in local recurrence rates translated into differences in 15-year mortality . In the decision for which women to offer preoperative MRI, both the absolute risk of future recurrence and the increased ac-curacy of MRI should be balanced. We believe that the use of pre-surgical MRI will become standard procedure, especially for younger women, for lobular cancers and in cases of mammography-ultrasound discrepancy in accordance with current guidelines from EUSOMA .
Neoadjuvant therapy will be continuously guided by image-based functional biomarkers
There are advantages with neoadjuvant therapy: the algo-rithmic choice of therapy can be evaluated in terms how it affects the actual tumor, and increase post-surgical compliance with adjuvant treatment. The tumor response can be complete (no tu-mor remains), partial, stable or progressive. To optimize treatment, it is helpful to be able to predict the final response at an early stage. MRI has been shown to be a powerful tool for response prediction [33,34]. If the final response is predicted to be poor the patient could be offered a change to a different treatment regimen. If a complete response is predicted, the planned surgery could be expedited. We believe that we will see new adaptive algorithms in the next years that makes use of biomarkers from several domains: imaging, pathology and genetics. Cross-discipline AI software might be applied to combine biomarkers from these various fields in ways that human specialists are unable to do. The use of MRI in the neoadjuvant setting is expected to increase.