Breast CA is a heterogeneous dz and rates of pCR after NAT range from 0.3-50.3% depending on tumor subtype.

Differences among individual tumor subtype occur bc of underlying variability in gene expression. Immunohistochemical markers (ER, PR) positivity and HER2 overexpression are used to classify tumors into 4 subtypes:

• (1) Triple-negative tumors (aggressive)

  • More chemosensitive bc of high cellular proliferation and commonly treated w NAT even in early stage.

  • pCR of 23.2%

• (2) HER2-positive tumors (aggressive)

  • More chemosensitive bc of high cellular proliferation and commonly treated w NAT even in early stage.

  • Additionally treated w HER2-targeted monoclonal antibody therapies.

  • pCR of 38.7%

• (3)ER/PR positive tumors - luminal A (less aggressive)

  • Lowest rates of pCR (0.3%)

• (4) ER/PR positive tumors - luminal B

  • More likely to be higher grade w higher markers of proliferation and less favorable prognosis.

  • Less likely to be treated w NAT as these tumors have the lowest rate of pCR bc of lower chemosensitivity.

  • Low rate of pCR (8.3%)

Interpretation of posttreatment MR images on changes in (1) tumor size and/or (2) enhancement compared w/ pretreatment MR images.

The authors' institution protocol performs a precontrast and 3 postcontrast axial T1WI sequences

• 1st post con: 90 sec

• last post con: 6 min

  • Most helpful in identifying residual dz, as enhancement may be delayed bc of the antiangiogenic effect of chemotherapy.

Either diameter or volume measurements maybe used to evaluate changes in tumor size.

• Changes in volume have been found to be more accurate in predicting response than changes in longest diameter

RECIST 1.1 guidlines specify that MRI is the preferred imaging modality to follow breast lesions in the neoadjuvant setting but is not used in axillary LN evaluation.

4 categories of response:

• Complete response: pCR in 19%

• Partial response: pCR in 45%

• Stable disease: pCR in 17%

• Progressive disease: pCR in 20%

In multifocal/multicentric breast CA, up to 2 lesions are measured in each breast, preferably the largest lesions, and are considered the target lesions that are to be measured at FU imaging. Any additional enhancing lesions in the breast are nontarget lesions and should be evaluated at FU imaging but do not require measurement.

Functional tumor volume (FTV) help predict recurrence-free survival.

• FTV incorporates enhancement thresholds to determine which portions of the tumor to include in a measurement, combining functional criteria w a volume measurement. 

• Visual assessment of decreased enhancement also indicates response to therapy, as effective chemotherapy reduces tumor neoangiogenesis.

Authors' institution: 

• Report tumor size by measuring diameters of the enhancing lesion in 3 dimensions without calculating tumor volume.

• No residual enhancement =  complete response 

• If residual enhancement -> measure enhancing lesion and compare to pretreatment. If the extent and degree of enhancement...

  • ...is clearly stable = nonresponse

  • ...has clearly decreased = partial response

• For more subtle/equivocal responses, follow RECIST 1.1 criteria and require at least a 30% decrease in diameter of the lesion to report partial response. In cases where diameter of lesion is unchanged but degree of enhancement is decreased, we report the decreased enhancement and note that it may represent partial response to therapy, as tumor cellularity may decrease without change in overall size of tumor.

Almost all patients proceed to surgery after NAT, regardless of whether MRI helps predict pCR. Therefore, the use of MRI in practice is not to help distinguish pCR from small-volume residual disease, but to assist in surgical planning. MRI can miss and overestimate residual disease. 

Overstimating residual dz can result in more extensive surgery; mastectomy instead of breast conservation and axillary dissection instead of sentinel LN biopsy. Missing residual dz may result in positive margins and the need for surgical reexcision.

Nonmass lesions are more likely than masses to result in a false-negative MRI in which there is complete imaging response to therapy but residual tumor is identified in the pathologic specimen.

Nonmass lesions have greater size discrepancies b/n MRI and pathologic analysis than do mass lesions. 

Given that ILC is most likely to manifest as a nonmass lesion, it is not surpirising that ILC is more likely to yield a false-negative MRI after neoadjuvant therapy than is IDC.

ILC and mixed ductal and lobular invasive CA also have larger discrepancies b/n size at MRI and pathologic analysis compared w/ IDC.

Growth pattern of ILC, characterized by diffuseness, multicentricity, and loss of cell-cell adhesion, may account for the lower accuracy of post-tx MRI in this histologic tumor type. 

Luminal (hormone receptor positive, HER2 neg) tumors are also more likely to be underestimated at posttreatment MRI. Although most commonly manifesting as a nonmass or diffuse lesion, as compared w triple-negative and HER2 postiive tumors, which usually manifest as discrete masses. It is important for surgeons and radiologists to know that there are larger size discrepancies b/n imaging and pathologic analysis and a higher change for positive margins after breast conservative therapy for hormon receptr positive and HER2 negative tumors.

Tumors responding to NAT may demonstrate different shrinkage patterns:

• (1) Complete response

• (2) Concentric shrinkage: 

  • A) Pattern in which pretherapy MR demonstrates a mass w/o surrounding NME or foci. 

  • B) Posttherapy MRI demonstrates a reduction in the longest diameter of the mass, +/- residual foci around the dominant mass. 

• (3) Nonconcentric shrinkage:

  • Any other pattern of shrinkage; crumbling pattern or multinodular lesion

  • Difficult to determine whether residual enhancement represents invasive cancer, DCIS, or reactive change after therapy 

  • Both over- and underestimation of residual dz is possible.

    • MR images may appear falsely negative in the setting of small scattered residual dz, as residual tumor cells may be too small to be identified at MRI or may be mistaken for benign foci.

Shrinkage patterns are associated w tumor subtypes.

• HER2 positive and triple negative tumors more commonly manifest as masses that demonstrate concentric shrinkage after NAT.

• Luminal tumors more commonly manifest as NME that demonstrate nonconcentric shrinkage.

  • Shrinkage pattern in luminal tumors is associated w prognosis 

    • Patients w luminal tumors that demonstrate concentric shrinkage have improved rates of survival compared w those w tumors that demonstrate noncencentric shrinkage, despite similarly low pCR rates.

Given the variable rates of response, baseline morphology and shrinkage patterns among the tumors subtypes, it is not suprising that the rates of FP and FN MRI studies vary but tumor subtype. Accuracy of breast MRI for predicting pCR is highest in triple negative and HER2 positive CA. 

NAT can be used to downstage the extent of dz in the breast to enable BCT and similarly in patients w known axillary LN mets, NAT can be used to downstage the axilla, with pCR rates of 35-68%. 

Sensitivity of post-NAT MRI to depict persistent LN mets is moderate, with studies reporting 61-72% sensitivity. 

A study of patients w abnormal LNs at pretreatment MRI, and subsequently with normal-appearing LNs at posttreatment MRI, found that 32% had metastatic disease in the axilla at surgery.

Specificity of MRI is also moderate. Studies of patients w abnormal LNs at posttreatment MRI found that only 58% were positive for LN mets at surgery.

Abnormal LNs at MRI appear distinct from other visible axillary LNS, including those in the contralateral breast. ANy of the following findings is abnormal:

• Cortical thickening

• Loss of the fatty hilum

• Round shape

• Irregular margin

• Heterogeneous cortex

• Surrounding edema

LNs can be considered normal when they are symmetric, are homogeneously enhancing, and have a thin cortex and preserved fatty hilum.

In patients who are not treated w NAT, a preoperative bx provden dx of mets to an axillary LN traditionally entailed LN dissection at the time of surgery (excision of levels I and II). Though studies showed that axillary LN dissection was not shown to improve survival or local control compared w sentinel LN bx for patients w one or two LNs w/ macrometastases and stage T1 or T2 primary tumor.

In sentinel LN bx, only tumor draining LNs are identified and excised (usually 1 to 4 LNs) making this procedure less morbid than axillary LN dissection which carries the risk of lymphedema, seroma, paresthesias