Translate RCT Design into Digital Twin Framework Randomized Controlled Trial (RCT)

The randomized controlled trial (RCT) (Fig (1)) is a gold standard for evaluating the effectiveness of FIT screening for colorectal cancer, aiming to reduce CRC incidence and mortality.

As the progression from subclinical to clinical state is often occult due to the interrupted medical regime it is therefore important to translate the RCT framework into the digital twin model that combines the physical entity with the virtual entity (in dotted line and circle) to reconstruct the natural history of CRC progression as shown in Fig(2).

Fig(3) adds basic screening characteristics including the uptake of FIT (attendance rate), FIT positive rate, and the referral rate of colonoscopy.

In real world data, it would be expected to yield different detection modes as a result of these basic screening characteristics. They consist of screen-detected CRC, post-colonoscopy CRC (PCCRC), CRC of non-compliance with colonoscopy, and FIT interval CRC.

The parameters used for generating the digital twin data are learned by applying the machine learning algorithm to the real world data.

Fig(4) is the extension of Fig(3) with the incorporation of advanced adenoma (AA).

Metaverse data is formed by the digital thread cohort that combines physical data with the virtual data.

Note that the digital thread data features each instance of the disease state (including advanced adenoma, subclinical CRC (SD-CRC), and clinical CRC) of the full natural history of CRC. Figures (2) and (3) show the details of physical data with the partial digital thread and virtual data with complete digital thread regarding the natural history of CRC progression in question.

Metaverse data will be stored in cloud with block chain technology. It enables one to create different of Avtar and facilitate the extended reality of metaverse consisting of augmented reality (AR) and mixed reality (MR).

The selection of Avtar is the core of the digital twin technology. Fig. (1) show the extraction of the partial natural history from the virtual group in commensuration with the physical detection mode.

In this example, occurrence of advanced adenoma (AA) among susceptible population is observed. To match the physical data mode, the candidates of phantom of the virtual group are extracted. Fig.(2) shows how to select the Avtar by matching the IAF among the candidate phantoms.

Augmented Reality (AR) for synthesizing the digital thread of colorectal neoplasia natural history of CRC progression.

Augmented reality (AR) is one of core technologies in the gamification of Metaverse using haptic glove, glass etc… to enjoy unbelievable immerse experiences from the virtual world.

Figure shows how to apply the concept of AR to create a series of the Avtar of the digital twins pairs as shown in the digital framework in the section earlier. These include Three steps are adopted. The first step is to provide a set of physical data ranging from the control group or non-participant, the screen-detected, and interval cancer. The second is to form metaverse data by combining the physical data with the virtual data.

The third step is to demonstrate how to identify an optimal Avtar after synthesizing the natural history of CRC progression from metaverse data given the designated patient’s information. The virtual data that makes significant contribution to identifying each Avtar are distinct.

The Avtar of the screen-detected group is synthesized by bridging the physical data on the occurrence of CRC (susceptible to subclinical state) with the virtual data on the progression from sub-clinical to clinical state that is hardly available from the real world data. The interval cancer is to overstretch the direct physical data from FIT negative to clinically-detected (CD) state into three-state processes with the synthesis of two jumping models:

FIT(-) → subclinical and subclinical → clinical state.

It is very important to see age at first screen or inter-screening interval, and time since negative screen until the occurrence of interval cancer are the main driver for identifying an optimal Avtar.

In addition, personalized information is also provided for identifying the Avtar. The similar synthesis is also made for overstretching the control group or non-participant into three-state process.

In 3-D dimension of the metaverse, the mixed reality (MR) is to change the object of the virtual world to see what would happen after the alteration.

The application of the mixed reality to colorectal cancer screening is like the provision of precision screening strategy to predict its effectiveness compared with the conventional screening strategy.

Figure(1)-(5) show how the concept of MR can be applied to screening scenario, taking FIT screening for example.

Figure show the framework of the digital twin model for applications (service layer) at population and individual level using both AR and MR (technique layer) in conjunction with and metaverse data (data layer).

Digital Twin Model for Evaluating Overdiagnosis of CRC and Adenoma

The first application is the use of the digital twin model to evaluating CRC and adenoma over-detection as a result of screening.

A series of the digital twin framework are demonstrated and applied to Taiwan national colorectal cancer screening data. The results show there are merely 10% and 16% overdiagnosis of CRC with and without considering the removal of adenoma.

While the above-mentioned f-Hb-guided precision screening strategy is adopted, the over-detection of adenoma is reduced to 12% compared with biennial screening strategy.

Digital Twin Model for Evaluating Care Cascade Colorectal Cancer

In metaverse health care, the developed deep machine multistage risk score (Delimited-RS) model mentioned earlier is applied to yield personalized trajectory of colorectal neoplasia from AA, SD-CRC, until CD-CRC based on three set of personalized features (IAF, MTF, and PPF).

The results of these projections are provided in Figure on the basis of Taiwan community-based integrated screening data. The N to 1 trial effectiveness of reducing CRC incidence and mortality in the light of various combinations of prevention, screening, and treatment) can be evaluated under the framework of digital twin model for metaverse health care in prediction and consultation when share-decision making is conducted beforehand.