麻豆传媒

Deep submucosal invasion (DSI) is considered a key risk factor for lymph node metastasis (LNM) and is therefore an important criterion for when to recommend surgery in T1 colorectal cancer (CRC). The risk for LNM ranges from 1% to 34%, depending on the presence of several known histologic risk factors.

To determine whether DSI is an independent risk factor for LNM, Barbara Bastiaansen, MD, of Amsterdam University Medical Center (AUMC) in the Netherlands, and colleagues conducted a meta-analysis, the results of which were.

Bastiaansen discussed the study in the following interview with the Reading Room.

The clinical background to this meta-analysis appears quite complex. Could you break it down?

Bastiaansen: Thanks to screening programs, the incidence of T1 CRC is rising, but only 11.2% of cases have LNM, meaning that the majority can potentially be cured by a local endoscopic resection. Surgical over-treatment remains a major issue in T1 disease, and optimal management must strike the right balance between maximizing oncologic safety and minimizing treatment-associated morbidity and mortality.

Traditionally, DSI has been regarded as a strong risk factor for LNM. Therefore, any visual signs of DSI raise red flags regarding the efficacy of endoscopic resection. Another consideration is that deeper invasion can hinder adequate lifting and thus increases the risk of perforation or incomplete endoscopic resection.

When optical classification systems were developed to differentiate superficial cancer from DSI, only superficial submucosal invasive T1 cancers were considered suitable for endoscopic resection. This has shaped T1 CRC management, as the majority of all cases have DSI at diagnosis.

Fortunately, recent advances in endoscopic resection techniques allow for the safe complete resection of T1 CRC even when deeper invasion is present. So, in order to optimize T1 CRC management and increase the chances of organ preservation, we felt the need to study the depth of submucosal invasion as an independent risk factor for LNM.

Did you have a working hypothesis about what the data would reveal?

Bastiaansen: Our hypothesis was that if we took into account the effect of all other risk factors, depth of invasion would lose significance as an LNM risk factor. Evidence from cohort studies had suggested that in the absence of other histologic high-risk features, DSI is associated with very low risk of LNM.

Although previous meta-analyses had designated DSI as an independent risk factor for LNM, these studies did not sufficiently correct for the confounding effect of concomitant established histologic risk factors, such as lymphovascular invasion, tumor budding, or poor differentiation grade.

A further question is whether it's really the depth of tumor invasion that significantly predicts LNM risk or whether this risk is more likely determined by actual tumor biology.

How did you select the studies for inclusion?

Bastiaansen: We initially assessed all studies reporting on the association between LNM and known histologic risk factors in patients with T1 disease who were treated with primary endoscopic resection, primary radical surgery, or completion surgery from 2000 to July 2021. Our final analysis included 67 eligible studies comprising more than 21,000 patients.

What were the principal findings and have they begun to impact practice at your medical center?

Bastiaansen: Our meta-analysis showed that DSI is not an independent risk factor for LNM. In fact, the absolute risk for LNM in T1 CRC with DSI as the sole risk factor was only 2.6%, obviating the need for radical surgery.

Here in the Netherlands, the national guideline on CRC management does not consider depth of invasion as a risk factor for LNM in T1 disease. So, at AUMC, all suspected cases of T1 disease suitable for radical local excision already undergo an endoscopic diagnostic excision as first-line treatment, regardless of the expected invasion depth. This policy is further strengthened by the results of this meta-analysis.

Did any findings surprise you?

Bastiaansen: Our study also showed that the proportion of all DSI cancers lacking other histologic high-risk criteria was almost 40%.

Has your group experienced any negative feedback because of this challenge to surgical orthodoxy?

Bastiaansen: No. Our multidisciplinary team works closely with our surgeons, who are very determined to deliver the best possible care for our patients. As pioneers in minimally invasive surgical care, they encourage and fully support endoscopic therapeutic development.

What kinds of cost savings might be expected from not using DSI as a sole indicator for first-line local excision?

Bastiaansen: Although not all DSI colon cancers will be suitable for a safe and complete diagnostic local excision, we expect that a significant proportion of all T1 disease can be completely removed with recent advanced resection methods. The number of curative resections will increase if DSI is disregarded as a risk factor for LNM, and this may lead to decreased morbidity, mortality, and medical costs.

What are the considerations going forward?

Bastiaansen: We need to be aware that even an oncologic surgical resection for T1 CRC will always carry a risk for disease recurrence of 2% to 5%. So, it seems unjust to simply compare surgical mortality directly with the risk for LNM to decide whether the patient needs an oncologic resection.

Furthermore, surgery has a more lasting impact on patients than can be determined in the 30-day follow-up of most studies. For example, the 5% to 7% post-surgical risk of anastomotic leakage and mortality after ileostomy reversal are not negligible.

Future research is needed to determine which patients truly have high-risk disease to spare others from radical surgery -- considering, of course, factors such as age, comorbidities, and patient preference.

What's the overarching take-home message for physicians treating early CRC?

Bastiaansen: It is very important that the case of every patient with suspected early CRC is discussed in a multidisciplinary team setting to determine, together with the patient, optimal treatment.

You can read the abstract of the study here.