Rectal cancer response assessment

Assessment of rectal cancer response to therapy,which may be chemotherapy, radiotherapy, or a combination,relies on the synthesis of clinical, endoscopic and radiologic evaluation. 

The purpose of neoadjuvant therapy is to downstage the tumour, to facilitate surgical resection, and reduce local recurrence rates. Up to 30% of patients undergoing chemoradiotherapy treatment may develop a complete response, so there is growing interest in the utility of MRI to identify this group of patients, in order to avoid surgery.

ChoiceThe choice of neoadjuvant therapy varies by geographic location, and the preferences of local units. Broadly speaking, in the United States, patients with tumours staged as T3 or above, and tumours with any positive nodes, will be offered neoadjuvant chemoradiotherapy in a combination known as long-course chemoradiotherapy (LCCR). In Europe, practice is usually more conservative with patients with early T3 tumours (T3a or T3b) being offered surgery, although node-positive patients with T3a or T3b tumours may still be directed to chemoradiotherapy.

Restaging MRI takes place between 6 and 10 weeks following completion of chemoradiotherapy. The key reporting parameters described below may form the basis of a template report.

Radiographic features

MRI

The images are compared to the baseline study, and an assessment made of any change in the volume and signal intensity of the primary tumour, lymph nodes, tumour deposits and extramural venous invasion, as well as the relationship of the tumour to adjacent structures such as the circumferential resection margin. The protocol used for imaging in this scenario is similar to primary staging, although diffusion weighted imaging plays a more important role here.

The typical response to therapy is of a reduction in tumour volume, with the development of fibrosis, manifestwhich manifests as intensely low signal on T2 weighted imaging. A complete response is identified by normalisation of the rectal wall, or the presence of linear low signal at the site of the treated tumour, with no restricted diffusion. The split scar sign is reported as having a high specificity and positive predictive value for a complete response ⁴. Some authors favour the use of 'near complete'"near-complete" as microscopic islands of tumour tissue may be present within the dense fibrosis and not detectable by MRI.

The presence of intermediate signal tissue, and foci of restricted diffusion, indicateindicates tumour that has only partially responded.

Treated circumferential tumours, even when exhibiting low signal on T2 weighted imaging and no restricted diffusion, are said to have a higher rate of residual tumour, whereas these signs can more reliably predict a complete/near complete-complete response with polypoid or semi-annular tumours ¹¹.

An atypical response is mucinous degeneration, with pools of high signal mucin within the tumour. In this setting it is not possible to state conclusively whether there is residual tumour tissue.

FindingThe presence of signs of complete response on all three of: endoscopy, T2 weighted imaging, and DWI lead to, indicates a true complete response rate of 98% ⁵.

With T4b tumours (invasion of other organs), even with the presence of fibrosis, the surgical approach will take into account which organs were invaded on the baseline study and remove them all, due to the risk of islands of microscopic residual tumour tissue within the fibrosis rendering an incomplete resection if a standard approach is used. Similarly, regression of tumour from the circumferential margin may be difficult to determine with certainty.

Up to 44% of nodes identified on the baseline study resolve completely ⁶. The same hurdles that apply to accurate determination of nodal status on baseline imaging apply to post-treatment MRI. Morphology is difficult to evaluate given that nodes shrink with therapy. The negative predictive value for nodal restaging is high due to the sterilising effect of chemoradiotherapy. However, one study found that identifying no nodes in the mesorectal compartment on the diffusion-weighted imaging is predictive of yN0 status ⁷,although this study only included 10 patients.

With regards to pelvic sidewall nodes, one study indicates that a reduction in size from 7 mm or more on pre-treatment imaging to 4 mm or less on restaging MRI was associated with markedly reduced pelvic sidewall recurrence, and a pelvic sidewall nodal dissection should be considered to achieve clearance if this criterion isn't met.

Oedema may be present within the mesorectal fat, a response to radiotherapy. A small volume of peritoneal fluid may also be seen.

On restaging MRI studies, the stage is prefixed with 'y'"y", for example, yT3a.

Tumour regression grade

For a few years, the tumour response was recommended to be graded using the MR tumour regression grade (mrTRG) ². This was the correlate of the pathologic tumour regression grade (pTRG). mrTRG grade 1 corresponds to a complete response, with only fibrosis detected, and mrTRG grade 5 corresponds to no response to therapy, with only tumour signal tissue identified.

However, in more recent times, the use of mrTRG has been largely discontinued, with data indicating a poor agreement between mrTRG and pTRG ³. In order to define the next steps in the patient journey, a pragmatic choice of three response categories is favoured; near/complete response, partial response, and poor response. These three categories enable three separate pathways to be followed.

For near/complete response, patients may be offered surgery, or may be entered into a 'watch"watch-and-wait' programme" program of active surveillance with the hope of avoiding surgery if there is a sustained clinical and radiologic response to therapy. For partial responders, as long as the anatomy allows, surgery will be offered. For poor responders, surgery may be offered if the anatomy allows, or further chemotherapy in an attempt to downstage the tumour.

For patients who have undergone a complete response, the 2-year tumour regrowth rate is 25%; the vast majority (97%) are on the luminal side, and the follow-up programmeprogram includes frequent MRI studies as well as clinical and endoscopic evaluation ⁸.

Pitfalls

The main pitfalls arise with diffusion-weighted imaging. High signal on the DWI images should be cross-referenced with the ADC map to ensure that it is true restricted diffusion rather than T2 shine through. Low signal on the ADC map should be cross-referenced with the DWI images before being interpreted as residual tumour; high signal on the DWI images indicate true restricted diffusion and therefore likely residual tumour, whereas low signal on the DWI and T2 images indicate fibrosis. Signal change in a star shape is likely to be due to luminal fluid, whereas focal or U shaped signal change is more likely to be tumour related ¹. DWI is not useful for the assessment of mucinous tumours due to the inherent high T2 signal of mucin. Finally, susceptibility artefact may occur due to gas within the lumen causing misleading signal change at the interface with the rectal wall, and research is underway to investigate whether microenemas may reduce this artefact.

Template report

Suggested items to include in a template report, based on the above pieces of data, are ^{12, 13,13}:

• change in volume of tumour and/or involvement of organs

• change in signal intensity of tumour (including split scar sign/fibrosis / fibrosis)

• appearances on diffusion weighted imaging

• length of residual tumour

• shortest distance to mesorectal fascia or levator muscle

• for low cancers, involvement of anal sphincter muscles or intersphincteric fat

• change in mesorectal and extramesorectal lymph nodes

• change in EMVI and tumour deposits

• category of response (complete/near complete-complete, partial or poor)

See also

• rectal cancer

• rectal MRI protocol

• rectal cancer staging

• approach to rectal MRI

• terminology of rectal cancer staging