A Digital Twin of RAS Wildtype Metastatic Colorectal Cancer to Evaluate the Efficacy of Standard-of-Care

Abstract

Background: Digital twin (DT) technology can increase efficiency of clinical trials. DT cohorts of metastatic colorectal cancer (mCRC) patients can serve as virtual standard-of-care (SOC) arms for reducing time, cost, and effort.We evaluated the feasibility of constructing a DT cohort of left-sided RAS wild-type mCRC and assessed the efficacy of SOC.

Methods: DT cohort of mCRC patients was constructed using the Phesi Trial Accelerator™ platform (>120 million patients, >4,000 indications). We selected patients aged 20-80 years, with RAS wild-type, unresectable, advanced/recurrent left-sided mCRC, untreated, with Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0-1. Exclusion criteria: prior chemotherapy, median progression-free survival (mPFS) unavailable.The first-line SOC treatment used in the selected studies was chemotherapy+bevacuzumab or cetuximab/panitumumab. The outcomes were median progression-free survival (mPFS) (primary), and median overall survival (mOS), median objective response rate (mORR) (secondary).

Results: Of 120,912,121 patients in 407,783 cohorts, we selected 1,455 chemotherapy-naïve patients (11 cohorts; 6 studies) with mCRC and included them for DT construction. The median age of the DT cohort was 62.5 years; 33% were women; the ECOG PS was 0 in 66%. DT analysis showed that mPFS (according to the cohort) for left-sided RAS wild-type mCRC patients with first-line SOC therapy was 11.9 months. The median of mOS and mORR were 34.3 months and 74.1% respectively.

Conclusions: It is feasible to construct a DT arm of mCRC patients and obtain the efficacy of SOC regimens using prior clinical trials data. DT arms can potentially increase trial efficiency and reduce the cost.

Introduction

The World Health Organization (WHO) has categorized cancer as one of the most prominent causes of mortality accounting for about 10 million deaths in 2020, of which colorectal cancer (CRC) caused 916,000 deaths.¹ CRC is the third most common type of cancer² and has a significant health and economic impact.¹

CRC encompasses a diverse set of tumors that are modulated by several distinct molecular pathways, some of which impact an individual's susceptibility to cancer and some which drive prognosis response to treatment. Left-sided and right-sided CRCs differ in gene expression, associated genomic alterations, and methylation profiles.³ The left and right colons have different embryonic origins; the right colon originates from the midgut, while the left colon and rectum from the hindgut. This leads to diverse vascularization in the proximal and distal colons.⁴ Notably, right and left side tumors are prognostically distinct as well. The mitogen-associated protein kinase (MAPK) pathway is one of the most altered pathways in CRC. The deregulation and activation of this pathway, driven by proto-oncogene activation, promotes cell proliferation, survival, and metastasis .⁵ Activation of this pathway influences the effectiveness of and resistance to specific therapies. MAPK pathway mutations occur more in left-sided CRC compared to right-sided CRCs except for BRAF alterations which are more often seen in the right.⁵

RAS/RAF wild-type CRC is a subtype of CRC where the tumors do not have mutations in the RAS or RAF genes and the absence of alterations in these genes (wild-type) is significant because it can influence response to therapies, notably response to anti-EGFR therapies.⁶

Treatment decisions for metastatic CRC (mCRC) depend on factors such as the patient's clinical status, tumor location and the status of predictive biomarkers: RAS (KRAS/NRAS), BRAF, and MSI. As a first-line therapy, epidermal growth factor receptor (EGFR) antibodies such as cetuximab and panitumumab can be used in combination with cytotoxic chemotherapy (FOLFOX or FOLFIRI) in patients with mCRC without RAS/BRAF mutations and left-sided tumors. A VEGF inhibitor, bevacizumab,may be used in RAS/RAF mutant tumors or right sided tumors in the 1^st line.⁷ Cetuximab, bevacizumab, and panitumumab, as adjunct to chemotherapy, have led to significant improvements in objective response rate (ORR), progression free survival (PFS), and overall survival (OS) in such mCRC cases.⁸ Left-sided RAS wild-type tumors have better clinical outcomes (ORR, PFS, OS) when treated with anti-EGFR antibodies plus chemotherapy compared to chemotherapy used alone or in combination with bevacizumab.^9,10

In the last few years, clinical research has seen several promising technological advances, bringing efficiency to conventional processes. Digital twin (DT) technology represents a step forward, towards the creation and use of digital replicas or simulations of real-world clinical trial processes, systems, and patient populations.¹¹ DTs are typically powered by artificial intelligence (AI), machine learning (ML), and data analytics.

Even in vigorously designed clinical trials with the same inclusion/exclusion criteria, the patient populations can vary widely in important parameters such as age, sex, body mass index (BMI), and so on. This is particularly true when sample sizes are small, as is usually the case in Phase I and Phase II oncology trials. DTs can help interpret the results from these trials more objectively and accurately.

DTs can simulate the characteristics of the target patient population for a clinical trial, helping stakeholders address challenges in recruitment, diversity, and optimize inclusion and exclusion criteria.¹² Researchers can simulate various scenarios to optimize trial designs by creating digital replicas of protocols and patient pathways, using predictive analytics. This includes determining efficient dosing regimens, identifying risks, and estimating trial duration.¹³

DT technology can help address ethical concerns around placebo arms; data from prior trials can be used to construct DTs which can function as virtual placebo/comparator groups. Virtual standard of care (SOC) trial arms can be constructed by leveraging already available clinical trial data reporting on the efficacy of the SOC regimens. This will translate to improvements in efficiency and costs.

In this study, we have used the Phesi Trial Accelerator™ platform (includes data from >70 million patients; >485,000 cohorts; covers >4,000 disease indications) and constructed a DT cohort of patients with left-sided mCRC and wild-type RAS. This platform has been previously used in other studies for efficacy evaluations of different drugs to treat other types of cancers.^14,15 Using the DT cohort as the standard of care (SOC) arm, we evaluated the efficacy of the first-line treatment of chemotherapy-naïve patients with RAS wild-type, unresectable, advanced/recurrent left-sided mCRC. Accordingly, the primary objectives of this study were:

1. to create a DT cohort with eligible mCRC patients and examine patient characteristics; and
2. to evaluate progression-free survival (PFS) value in the DT cohort.

The secondary objectives were to examine the overall survival (OS) and objective response rate (ORR) in the DT cohort.The aim was to validate the DT cohort by comparing its predictions with real-world clinical outcomes from the literature. This study hopes to add evidence supporting DT technology for designing, conducting, and interpreting clinical trials in the future.

Methodology

Construction of the DT cohort

A DT cohort was constructed using the Trial Accelerator™ platform. This platform has been described in our previous work.¹⁵

Patient characteristics, demographics, concurrent medications, comorbidities, disease status, and treatment outcomes, were identified using an AI algorithm. The AI algorithm was then applied to identify patient data that are in line with efficacy outcome measures, and patient inclusion and exclusion criteria. The final patient data were put through a detailed quality analysis and verification process utilizing manual review to ensure that the DT cohort conformed to the inclusion and exclusion criteria.

This study did not enroll patients from healthcare facilities, but sourced patient data from published studies, therefore an ethics committee approval was not required.

Inclusion criteria (studies including patients as per following criteria): patients aged ≥20 and <80 years at the time of informed consent, who have RAS wild-type, unresectable, advanced/recurrent left-sided metastatic CRC, are chemotherapy naïve, and have Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0 to 1; studies should report the median PFS (mPFS) value.

Exclusion criteria: Studies were excluded from the analysis if they included patients with prior chemotherapy, and if they did not report the mPFS value.

Interventions

All the patients received first-line standard-of-care (SOC) treatment including chemotherapy (FOLFIRI/FOLFOX/FOLFOXIRI) with either anti-epidermal growth factor receptor inhibitor (anti-EGFR) (cetuximab or panitumumab) or anti-vascular endothelial growth factor inhibitor (anti-VEGF) (bevacizumab) therapy.

Patient characteristics, mPFS, mOS, and mORR evaluation

The patient characteristics in the DT cohort such as age, sex, ECOG Performance Status (ECOG PS), site of metastasis, number of metastases, and RAS and BRAF mutation status were examined and analyzed. The mPFS values were extracted and analyzed as well. Similarly, the median OS (mOS), and median ORR (mORR) values from the studies that reported these parameters were extracted and evaluated.

Statistical analysis

Statistical analysis was performed using Excel software, and descriptive statistics based on central tendency parameters are presented for each outcome measure. Weighted median values of the mPFS, mOS, and mORR values were calculated based on the number of patients in each cohort. These weighted median values of mPFS, mOS, and mORR are reported as the predicted values for mCRC left-sided tumors with wild type RAS status treated with SOC. Categorical variables are provided as percentages, and continuous variables as median values.

Results

DT cohort

The patient selection flow diagram is presented in Figure 1. Of the 120,912,121 patients in 407,783 cohorts represented in the Trial Accelerator^TM database, 226,398 mCRC patients from 1,207 cohorts were identified; from these, 3,255 mCRC patients with left-sided RAS wild-type mCRC were selected from 22 cohorts; further, from these, 1,455 patients from 11 cohorts who were treatment-naïve and with available mPFS data were selected and included in the construction of the DT cohort.

Patients were included from trials conducted globally, including CAIRO5, Valentino, MACRO-2, PLANET, PRIME, PEAK, and APEC studies (additionally, studies by Köhne et al.¹⁶ and Watanabe et al.¹⁷ were also included) (eTable 1 includes all details about the included studies). Data collection for the patients in the various studies occurred between the years 2009 and 2022.

Patient characteristics of the DT cohort

Table 1 shows the characteristics of the DT cohort. The median age of the patients was 62.5 years. One-third of the population (33%) were women, and two-thirds (67%) were men. The ECOG PS was 0 in a majority of the patients (66%). All included patients were RAS wild-type, and 96% were BRAF wild-type. The liver was the only site of metastasis in 27% of the patients.

The baseline and disease characteristics of the DT cohort are detailed in Table 1.

Efficacy analysis

The mPFS, mOS, and mORR from the 1,455 patients (in 11 cohorts) used to construct the DT cohort were analyzed. Manual checks were performed to ensure that the included studies reported the median values for these parameters. All the patients were chemotherapy-naïve and received first-line SOC treatment. The details of the components of the SOC regimens are given in eTable 2 for each of the included cohorts.

Primary objective

mPFS of the SOC DT cohort

Figure 2 shows the mPFS of the 11 cohorts used to construct the DT cohort. The mPFS values ranged from 11 months to 14.6 months. The mPFS values were quite stable across the cohorts, even though the size of individual cohorts varied (43 patients to 312 patients); additionally, the mPFS values remained quite stable even across cohorts that were evaluated over a broad time (year of data collection ranged from 2009 to 2022). The calculated weighted median mPFS value (weighted according to the size of the cohort) for the DT cohort was 11.9 months. Thus, the predicted mPFS for left-sided RAS wild-type mCRC patients after first-line SOC therapy is 11.9 months as per this DT analysis.

Secondary objectives

mOS in the SOC DT cohort

The mOS data were available for 7 out of the 11 cohorts selected for DT construction. The cohorts without mOS data (cohorts 1, 7, 8, and 9) were excluded from the mOS calculation. Figure 3 shows the mOS arranged by year of data collection in the 7 individual cohorts that reported these values. The mOS ranged from 30.6 months to 43.4 months. Our analysis showed that the calculated weighted median mOS (weighted according to the size of the cohort) was 34.3 months for the DT cohort. Thus, the predicted mOS for left-sided RAS wild-type mCRC patients after first-line SOC therapy is 34.3 months as per this DT analysis.

mORR in the DT cohort

The mORR data were available for 8 out of the 11 cohorts selected for DT construction. The cohorts without ORR data (cohorts 2, 3, and 4) were excluded from the ORR calculation. Figure 4 shows the mORR arranged by year of data collection in 8 individual cohorts that reported these values. The mORR ranged from 53% to 80.2%. The median ORR weighted according to the size of the cohort for the SOC DT was 74.1%. Thus, DT analysis predicts that the median ORR for left-sided RAS wild-type mCRC with SOC therapy will be 74.1%.

Discussion

Two major challenges that face clinical investigators conducting clinical trials and the pharmaceutical industry include the cost and time investment in patient recruitment, and this is compounded in the case of studies that need to satisfy a very specific set of inclusion criteria.¹⁵ AI-assisted computational modeling shows great promise as a complement to traditional clinical trial design and conduct methodologies and helps in overcoming the above-mentioned challenges; technologies such as DTs are being used to simulate both patients as well as diseases, and can help model the efficacy and safety of drugs.¹⁸ AI-assisted modeling of diseases and patient data is poised to take its place in drug discovery and the conduct of clinical trials and awaits regulatory approval. The FDA has included a section promoting education on Model-informed Drug Development (MIDD) in the Prescription Drug User Fee Act (PDUFA); MIDD includes support for AI-assisted disease and patient modeling. Special regulatory pathways are being established for newer AI-assisted drug discovery modalities;¹⁸ it is reasonable to expect that this support will extend to DTs as formally accepted clinical trial arms in the future.

In this study, we evaluated the feasibility of constructing a DT cohort of chemotherapy-naïve patients with left-sided wild-type RAS/BRAF mCRC treated with first-line SOC regimens using the Trial Accelerator™ platform. We also evaluated the efficacy outcomes of the SOC regimens in this cohort. Our results show that it is feasible to construct a DT cohort of mCRC patients using publicly available data from previously conducted clinical trials. The baseline characteristics of the DT cohort showed a median age of 62.5 years; there was a higher prevalence of men, in agreement with literature reports showing that left-sided mCRC is seen more in men while right-sided mCRC predominates in women.^19,20 Additionally, analysis of the DT data yielded predictions of efficacy outcomes of the first-line SOC regimens in terms of mPFS, mOS, and mORR. The weighted mPFS was 11.9 months, weighted mOS was 34.3 months, and the weighted mORR was 74.1% for the DT.

Tejpar et al. (2017), in a retrospective analysis of 149 patients from the CRYSTAL and FIRE-3 trials on left-sided RAS wild-type mCRC reported that after first-line therapy the mPFS was 12.0 months (FOLFIRI+cetuximab)in the CRYSTAL trial and 10.7 months in two arms (identical values for FOLFIRI+cetuximab and FOLFIRI+bevacizumab arms) in the FIRE-3 trial. Other studies have reported mPFS values of 13.3 months (Samalin et al., 2024), 11.2 months, and 12.7 months (CALGB 80405 trial, FOLFIRI/FOLFOX+bevacizumab and FOLFIRI/FOLFOX+cetuximab, respectively),¹⁹ and 15.0 months (PFS, cetuximab arm)²¹ for this patient population. The predicted value as per DT analysis was 11.9 months for mPFS for this patient profile in the current study, and this fits well in the context of the above-mentioned reports.

Published studies on left-sided RAS wild-type mCRC treated with first-line therapy have reported mOS values of 32.6 and 39.3 months (CALGB 80405 trial, FOLFIRI/FOLFOX+bevacizumab and FOLFIRI/FOLFOX+cetuximab, respectively),¹⁹ 37 months (chemotherapy+bevacuzumab),²² and 35.8 months (cetuximab).²¹ This is in close alignment with the mOS value derived with the DT analysis in this study, at 34.3 months. Similarly, the reported values for ORR include 57.9% and 69.4% (FOLFIRI+cetuximab and FOLFIRI+bevacizumab, respectively),¹⁹ 72.5%, 68.8%, and 61.1% (FOLFIRI+cetuximab, CRYSTAL and FIRE-3, and FOLFIRI+bevacizumab, FIRE-3, respectively).²⁰ Our calculated mORR as per DT analysis was 74.1%, which is also aligned with the literature-reported values.

In addition to agreement with literature reports, the efficacy values tended to be stable across the cohorts that were used to construct the DT, regardless of the size of the cohort (43 patients to 312 patients) or the year of data collection (spanning more than a decade, from 2009 to 2022). This speaks to the reliability and robustness of the data that were used to construct the DT, which in turn imparts reliability and robustness to the DT. The above discussion placing the DT results in the context of relevant literature shows the potential for DTs to effectively represent baseline and SOC cohorts and strengthens the argument for the inclusion of DT arms in clinical trials.

There are several limitations associated with the use of DTs. The biggest challenge to constructing DTs is data acquisition. In this study, however, the use of the Trial Accelerator™ platform provided a great advantage due to the high volume of quality data included in this database, and this is one strength of the current study. Another limitation of DTs and the current study as well is that a variety of studies conducted from different parts of the world are included for DT construction, and this can increase data heterogeneity. It can be noted that among the 11 included cohorts in the study, mOS and mORR data were not available for several cohorts. If clinical trial designs are optimized and standardized globally to ensure uniformity in outcome measurement and reporting of details, it would increase ease of DT construction. It would also help in the development and more extensive application of this promising technology. One more limitation is that a small proportion of patients (4%) showed BRAF mutations. While it is now accepted that patients with BRAF mutations should not be given anti-EGFR therapy, the data in question were collected in 2009, before the above became widely accepted. Additionally, the proportion of patients is very small, and the impact on the results may be small. Another point to be noted is that the generalizability of the study findings is limited, and the findings can be applied only to the patient populations that satisfy the inclusion criteria. Thus, it is important to construct further DTs with a broader set of inclusion criteria and validate such cohorts in larger populations.

DT technology has great potential for not only modeling patients and diseases but also across all stages of drug discovery.DTs can also enable real-time monitoring of patient data and trial progress. They can identify deviations from the planned course, allow timely interventions, reduce data errors, and improve data quality.¹⁵ Furthermore, by maintaining a digital record of all trials and data, DTs can help ensure regulatory compliance and facilitate the preparation of regulatory submissions.^23,24 DT technology can help in creating a diverse and more inclusive patient population for clinical trials and aid sponsors in complying with the draft FDA guidance on diversity and inclusion.

Conclusion

In summary, in this study we constructed a DT of mCRC patients with specific inclusion criteria using already available data. Thus, our study can be used as proof of concept for the feasibility of constructing a DT control arm from historical data. We show that this DT can be used to simulate SOC treatment and provide reliable outcomes. DTs can function as virtual SOC arms to improve cost effectiveness and efficiency of clinical trials. DTs can also help overcome ethical challenges associated with placebo/control arms. It is hoped that this study will add to the growing body of knowledge providing proof of concept for the potential integration of DT technology in all steps of drug discovery and development.

*Gen Li, PhD, MBA

Founder and President, Phesi

*Aparna R. Parikh, MD

Director, Global Cancer Care Program, Mass General Brigham Cancer Center, Massachusetts General Hospital

*Co-corresponding authors

Both authors contributed equally to the manuscript

Acknowledgments

The authors would like to acknowledge and thank Srividya Malkapuram, PhD (Pharmacology) (Affiliation: Krystelis) for providing medical writing support, which was funded by Phesi.

Author contributions

GL was responsible for developing the patient database, screening and finding potentially eligible studies, extracting and analyzing data, constructing the tables and figures, and interpreting results. ARP was responsible for guiding data interpretation and discussion, reviewing the draft, and providing input on manuscript content. GL and ARP performed source data verification and contributed toward manuscript drafting, reviewing, and approving the final version.

Data access, responsibility, and analysis

GL had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Funding support/role of funder

This study was funded by Phesi. The funder had a role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, and approval of the manuscript; and decision to submit the manuscript for publication.

Conflict of Interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: GL reports financial support and article publishing charges were provided by Phesi. GL reports a relationship with Phesi that includes: board membership, employment, and equity or stocks. ARP reports a relationship with Phesi that includes: consulting or advisory.

Data availability

The data underlying this article are available in the article and in its online supplementary material.

References

1. World Health Organization. Available at: https://www.who.int/news-room/fact-sheets/detail/cancer. Accessed 19 November 2024

2. Omran AR. The epidemiologic transition: a theory of the epidemiology of population change. 1971. Milbank Q. 2005;83(4):731-757. doi:10.1111/j.1468-0009.2005.00398.x

3. Hanna DL, Lenz HJ. How we treat left-sided vs right-sided colon cancer. Clin Adv Hematol Oncol. 2020;18(5):253-257. doi:10.1001/jamaoncol.2024.5356

4. Glebov OK, Rodriguez LM, Nakahara K, et al. Distinguishing right from left colon by the pattern of gene expression. Cancer Epidemiol Biomarkers Prev. 2003;12(8):755-762.

5. Bellio H, Fumet JD, Ghiringhelli F. Targeting BRAF and RAS in Colorectal Cancer. Cancers (Basel). 2021;13(9):2201. doi:10.3390/cancers13092201

6. Knijn N, van de Water C, van Vliet S, et al. Sequencing of RAS/RAF pathway genes in primary colorectal cancer and matched liver and lung metastases. Appl Cancer Res. 2019;39(9). doi:10.1186/s41241-019-0079-y

7. Lakatos G, Köhne CH, Bodoky G. Current therapy of advanced colorectal cancer according to RAS/RAF mutational status. Cancer Metastasis Rev. 2020; 39:1143–1157. doi:10.1007/s10555-020-09913-7

8. Douillard JY, Oliner KS, Siena S, et al. Panitumumab-FOLFOX4 treatment and RAS mutations in colorectal cancer. N Engl J Med. 2013;369(11):1023-1034. doi:10.1056/NEJMoa1305275

9. Aljehani MA, Morgan JW, Guthrie LA, et al. Association of Primary Tumor Site With Mortality in Patients Receiving Bevacizumab and Cetuximab for Metastatic Colorectal Cancer. JAMA Surg. 2018;153(1):60-67. doi:10.1001/jamasurg.2017.3466

10. Nevala-Plagemann C, Iyengar S, Trunk AD, Pappas L, Haaland B, Garrido-Laguna I. Treatment Trends and Clinical Outcomes of Left-Sided RAS/RAF Wild-Type Metastatic Colorectal Cancer in the United States. J Natl Compr Canc Netw. 2022;20(3):268-275. doi:10.6004/jnccn.2021.7079

11. Dihan MS, Akash AI, Tasneem Z, et al. Digital twin: Data exploration, architecture, implementation and future. Heliyon. 2024 Mar;10(5):e26503. doi:10.1016/j.heliyon.2024.e26503

12. Popowitz E. Digital twins are revolutionizing clinical trials and so much more. Healthcare Trends. 2024. (definitivehc.com) https://www.definitivehc.com/blog/digital-twins-clinical-trials#:~:text=In%20clinical%20trials%2C%20digital%20twins%20are%20virtual%20replicas,and%20organs%20designed%20to%20predict%20specific%20treatment%20outcomes. Accessed January 24, 2025.

13. Mitsi G, Grinnell T, Giordano S, et al. Implementing Digital Technologies in Clinical Trials: Lessons Learned. Innov Clin Neurosci. 2022;19(4-6):65-69.

14. Li G, et al. A Digital Twin on CAR-T Cytokine Release Syndrome (CRS) patients with Standard of Care measured by CRS distribution by grade. Applied Clinical Trials. 2023. https://www.appliedclinicaltrialsonline.com/view/a-digital-twin-on-car-t-cytokine-release-syndrome-crs-patients-with-standard-of-care-measured-by-crs-distribution-by-grade. Accessed January 24, 2025

15. Li G, Chen YB, Peachey J. Construction of a digital twin of chronic graft vs. host disease patients with standard of care. Bone Marrow Transplant. 2024;59(9):1280-1285. doi: 10.1038/s41409-024-02324-0

16. Köhne CH, Karthaus M, Mineur L, et al. Impact of Primary Tumour Location and Early Tumour Shrinkage on Outcomes in Patients with RAS Wild-Type Metastatic Colorectal Cancer Following First-Line FOLFIRI Plus Panitumumab. Drugs R D. 2019;19(3):267-275. doi:10.1007/s40268-019-0278-8

17. Watanabe J, Muro K, Shitara K, et al. Panitumumab vs Bevacizumab Added to Standard First-line Chemotherapy and Overall Survival Among Patients With RAS Wild-type, Left-Sided Metastatic Colorectal Cancer: A Randomized Clinical Trial. JAMA. 2023;329(15):1271-1282. doi:10.1001/jama.2023.4428

18. Moingeon P, Chenel M, Rousseau C, Voisin E, Guedj M. Virtual patients, digital twins and causal disease models: Paving the ground for in silico clinical trials. Drug Discov Today. 2023;28(7):103605. doi:10.1016/j.drudis.2023.103605

19. Arnold D, Lueza B, Douillard JY, et al. Prognostic and predictive value of primary tumour side in patients with RAS wild-type metastatic colorectal cancer treated with chemotherapy and EGFR directed antibodies in six randomized trials. Ann Oncol. 2017;28(8):1713-1729. doi:10.1093/annonc/mdx175

20. Tejpar S, Stintzing S, Ciardiello F, et al. Prognostic and Predictive Relevance of Primary Tumor Location in Patients With RAS Wild-Type Metastatic Colorectal Cancer: Retrospective Analyses of the CRYSTAL and FIRE-3 Trials. JAMA Oncol. 2017;3(2):194-201. doi:10.1001/jamaoncol.2016.3797

21. Lu HJ, Lin JK, Chen WS, et al. Primary tumor location is an important predictive factor for wild-type KRAS metastatic colon cancer treated with cetuximab as front-line bio-therapy. Asia Pac J Clin Oncol. 2016;12(3):207-215. doi:10.1111/ajco.12469

22. Snyder M, Bottiglieri S, Almhanna K. Impact of Primary Tumor Location on First-line Bevacizumab or Cetuximab in Metastatic Colorectal Cancer. Rev Recent Clin Trials. 2018;13(2):139-149. doi:10.2174/1574887113666180328104109

23. Grieb N, Schmierer L, Kim HU, et al. A digital twin model for evidence-based clinical decision support in multiple myeloma treatment. Front Digit Health. 2023;5:1324453. doi:10.3389/fdgth.2023.1324453

24. Food Drugs and Administration. Available at: https://www.fda.gov/media/177030/download?attachmenthttps://www.who.int/news-room/fact-sheets/detail/cancer. Accessed 19 November 2024