TL;DR — Since Taiwan’s Regenerative Medicine Acts came into force on 1 January 2026, three interlocking layers of institutional design have gradually taken shape: the dual Acts themselves, the Article 9 conditional approval (有附款許可, “approval with附款”) under the Regenerative Medicinal Products Act, and the Taiwan Regenerative Medicine Advanced Therapy Pilot (T-RMAT) jointly run by TFDA and CDE, announced in March 2026. Set against international peers, this design sits in the same policy trend as U.S. RMAT, EU PRIME, and Japan Sakigake. The point is not to lower review standards, but to allow products meeting specific conditions to access the market earlier — when risk-benefit data already shows initial support — and trade earlier access for post-market verification, risk management, and continued monitoring. Taiwan’s regulatory framework is largely in place. What matters next is whether industry actually files through this pathway and accumulates verifiable post-market cases.

I have been talking recently with some Japanese partners about a project: whether to bring a cell therapy that has completed Phase II in the U.S. into Taiwan. My instinct was that this path would still depend heavily on a full Phase III, additional local trials, and a long wait for approval. But once I started unpacking the system, I realized that Taiwan had been quietly laying a path that diverges from that traditional assumption over the last few years.

Behind this project are patients who have been waiting for a long time. Many regenerative medicine therapies have already shown meaningful efficacy at the U.S. Phase II stage, yet patients in Taiwan typically wait another 5 to 10 years — for Phase III to complete, for product launch, for entry into Taiwan, and for NHI reimbursement. Some patients do not make it. For diseases that progress quickly, every link in the chain that can be shortened becomes a real, concrete option.

In June 2024, the Legislative Yuan passed the Regenerative Medicine Act and the Regenerative Medicinal Products Act, promulgated by the President; both came into force on 1 January 2026. At the same time, TFDA and CDE rolled out the Taiwan Regenerative Medicine Advanced Therapy Pilot (T-RMAT), announced in March 2026, as a supporting layer for early regulatory dialogue, review planning, and pre-submission guidance. Stacked together, these mechanisms create — for therapies “with mature Phase II evidence, aimed at life-threatening or seriously disabling diseases” — a more discussable possibility of accelerated entry than ever before.

There are at least two common misreadings on this pathway. First, hearing “post-Phase II therapies could apply for conditional approval” makes many people think this means lowered review standards. Second, many multinational pharmas evaluating the Asian market still view Taiwan as a traditional follower of U.S. and EU data, and have yet to factor the new conditional approval and pilot scheme into the same strategic plan. In practice, both intuitions deserve a fresh look.

Why can a post-Phase II therapy apply for market entry? How the three layers stack

Returning to the Regenerative Medicinal Products Act itself, the conditional approval set up under Article 9 is the core of this pathway. Under the article, an applicant must simultaneously satisfy conditions concerning disease characteristics, clinical trial progress, and risk-benefit data; not every product that completes Phase II is automatically eligible.

First, the indication must qualify as a “life-threatening disease” or a “seriously disabling disease”. “Life-threatening” refers to a disease that has progressed to a stage where death could reasonably occur within months, or where untreated cases may result in premature death; the emphasis is on imminent mortality risk. “Seriously disabling” refers to a disease that has severely impaired bodily function, affected daily living, or rendered the patient unable to live independently, and which may further deteriorate without treatment. Transient or self-resolving sequelae or complications are not included.

Second, the applicant product must have completed Phase II human trials. In other words, the product is no longer at the proof-of-concept stage; it has accumulated safety and preliminary efficacy data in a meaningful number of subjects.

Third, the applicant must submit data sufficient to support a risk-benefit assessment, and the data must pass review by the competent authority as showing safety and preliminary efficacy, and then be cleared by the Regenerative Medicine Review Board, before conditional approval can be granted. The Board’s role is to bring medical, ethical, industrial, patient and review-practice perspectives to the same table, and decide whether a particular product is worth trading “earlier launch” for “subsequent verification obligations”. To put it plainly: this is not “Phase II finished, therefore on the market”, but “Phase II finished, therefore eligible to be rigorously discussed for possible conditional launch”.

Once conditional approval is granted, the sponsor receives a permit valid for up to five years, non-renewable. During this period, the sponsor must continue conducting confirmatory efficacy trials or provide real-world data (RWD) studies of comparable evidentiary strength, and submit data periodically to the competent authority. Failure to fulfill the附款 (附帶 conditions) or the emergence of major safety concerns may lead the competent authority to revoke the conditional approval.

So this ticket is not a blank check. It is more like a permit with a deadline, with attached obligations, and which can be withdrawn at any time. The first stage uses completed Phase II data, risk-benefit review and Board judgment to trade for earlier access; the second stage requires continued confirmatory efficacy trials and periodic data submission post-launch. If obligations are not honored, or if the competent authority assesses major safety concerns, the permit can be revoked. This is different from “lowering review standards”. The latter directly lowers the bar; the former shifts part of the evidence requirements into a staged completion.

The second layer is the underlying framework of the dual Acts. The Regenerative Medicine Act governs the institutional execution side — clinician qualifications, facility standards, human trials, informed consent, adverse event reporting. The Regenerative Medicinal Products Act governs the product side — from registration, conditional approval, manufacturing and distribution, post-market management, to pharmaceutical injury remedies — placed within a “total lifecycle management” framework comparable to that for conventional drugs. For sponsors, this means competitiveness is no longer just about science; it also includes PIC/S GMP, GDP, batch consistency, traceability, cold chain logistics, and long-term risk management capability.

The third layer is T-RMAT. TFDA and CDE announced T-RMAT in March 2026, positioned as an accelerated guidance mechanism aimed at letting eligible regenerative medicinal product sponsors build scientific consensus with the reviewers before formal submission, and align clinical, CMC (chemistry, manufacturing and controls), toxicology, and statistics data with review expectations from the start. The planned guidance model includes bi-weekly bilateral communication with both agencies, rolling review, written review opinions, and binding effect of early consultation conclusions when conditions remain unchanged. When sponsors eventually submit for conditional approval or BLA (Biologics License Application), the target review timeline can be compressed to 120 days; for IND (Investigational New Drug), the target is 15 days. These are timelines officially stated by the agencies; actual case outcomes still depend on the completeness of the submitted data.

In practice, T-RMAT prioritizes the following kinds of cases: regenerative medicinal products already covered by existing TFDA or CDE guidance and consultation tracks; products planning to file BLAs (including conditional approval) in 2026–2027; first-in-human trials for exosomes or extracellular vesicles (exosome / EV); and regenerative products of public-health value with high clinical urgency.

Stacked together, these are not three independent tools but a continuous design of “alignment at the scientific consultation stage → clear conditional approval threshold → post-market verification and safety monitoring linked together”.

Placing Taiwan back on the global accelerated-pathway map

This pathway is not isolated in the global coordinate system. The U.S. FDA’s RMAT designation (Regenerative Medicine Advanced Therapy) has been operating since the 21st Century Cures Act of 2016. According to FDA cumulative data as of 30 September 2025, CBER had received 388 RMAT designation requests, of which 193 were granted, 165 denied, and 14 withdrawn. The numbers make one point clear: RMAT is not a niche curiosity but a substantive accelerated pathway in the U.S. regenerative medicine review system. The EU EMA also launched PRIME (PRIority MEdicines) the same year, with early CHMP/CAT rapporteur appointment, iterative scientific advice, protocol assistance, and accelerated assessment alignment; in 2024 it received 58 requests and granted 15 designations.

The most instructive comparison for Taiwan, however, is Japan. Japan’s Sakigake designation, launched in 2015, has from the start combined “first developed in Japan”, “early clinical data showing prominent effectiveness”, “for unmet medical need”, and “prioritized consultation and accelerated review” within a single logic, shortening the review timeline from 12 months to 6 months. From 2017 onwards, Sakigake was further combined with the conditional approval pathway for regenerative medicinal products, forming Japan’s “front-end accelerated guidance, back-end conditional access, then post-market data to verify efficacy and safety” institutional combination. Taiwan’s T-RMAT plus conditional approval, while not identical in detail, follows a closely comparable policy logic. For therapies drawing on U.S., Japanese and European data, Taiwan now sits closer to “running alongside” than “running behind”, as it did over much of the last decade.

On the cross-border collaboration ground: why this path is missed

From the cross-border collaboration standpoint, what most often trips this system up is not design, but information lag.

Many overseas teams still understand Taiwan in terms of “small market, regulatory timing trailing U.S. and EU, limited entry incentives”. But in regenerative medicine, Taiwan’s institutional update speed over the past few years has actually been quick: dual Acts in force, conditional approval online, T-RMAT pilot announced, Regenerative Medicine Review Board and post-market management framework gradually taking shape. These shifts have not yet fully translated into the international industry’s shared awareness.

Another common misjudgment is treating “Phase II plus launch” as “lower evidence requirements”. In practice, the post-approval confirmatory efficacy obligations, post-market safety monitoring, source and distribution tracing, and pharmaceutical injury remedy designs, taken together, do not make conditional approval cheaper than full approval. The real difference on this pathway is not “cheaper”, but “saving time and letting patients in need access therapies earlier”. For therapies with patients actually waiting in Taiwan, that time difference can be the difference between making it or not.

A third observation worth bringing to the table is that this path also suits therapies with accumulated East Asian population data. Not because Taiwan only looks at East Asian data, but because regenerative medicinal products involve immune response, cellular dynamics, long-term safety, and process variation, making the question of whether foreign clinical data can be extrapolated to Taiwan harder than for conventional drugs. If a product already has interpretable clinical data in Japan, Korea, Taiwan, or other East Asian populations, reviewers can more easily discuss “what can be extrapolated, what needs further work”. Many therapies currently taking the Japan Sakigake-plus-conditional-approval route happen to meet this profile.

In the course of cross-domain collaboration in the circular economy and semiconductors, I have lived through several moments where “the rules had already changed, but the industry had not yet caught up”. In those moments there are usually two kinds of people: those who stay tied to existing impressions and continue on familiar but outdated paths, and those who go back and re-read the rules, then realign their external story and internal timeline. The latter tend to be the group that enters new markets fastest in the years that follow.

Not lowering review standards — moving part of the evidence requirement to staged completion

If we only remember “Phase II can launch”, we miss the other side of conditional approval.

First, conditional approval is valid for up to five years and is non-renewable. This is a strictly rigid condition. If the sponsor cannot complete subsequent verification, or cannot provide data sufficient to support transition to full approval within the statutory period, conditional approval cannot simply be maintained indefinitely.

Second, conditional approval does not exempt subsequent efficacy verification. By design, the sponsor must continue running confirmatory efficacy trials or provide real-world data studies of comparable evidentiary weight during the permit period, and submit data periodically to the competent authority. In practice, this is often best understood as: a portion of what would conventionally be confirmatory evidence completed pre-launch is now completed post-launch — but does not necessarily map one-to-one onto a traditional, narrowly defined Phase III.

Third, the applicant does not merely submit clinical data. Under the conditional approval regime, the sponsor must also submit an efficacy verification plan, fee and collection arrangements, patient remedy measures, and other matters designated by the competent authority. The system thus requires not just scientific evidence but also financial transparency and risk-bearing arrangements at the point where patients actually access the therapy.

Fourth, a regenerative medicinal product that obtains conditional approval does not escape general drug lifecycle management requirements. Based on the competent authority’s public statements and the relevant regulatory context, manufacturing quality, distribution management, source and flow tracking, post-market safety monitoring, and pharmaceutical injury remedies must still be implemented in parallel. For applicant companies, this means conditional approval is not “get the market entry first, do internal compliance later”; rather, a meaningful level of quality and supply-chain readiness must already be in place before entering this pathway.

Stacking the four points together makes the underlying design logic clearer. Conditional approval is not “approve first, sort it out later”. It is the competent authority, in conditions of specific disease, high medical need, and preliminary efficacy and safety support, trading earlier access for more intensive post-market verification, risk management and continued monitoring. For sponsors, this is not a reduction in responsibility but a shift — moving part of the responsibility from “pre-launch waiting” into ongoing post-launch regulatory obligations.

Conditional approval, Review Board oversight, the T-RMAT pilot, pharmaceutical injury remedies, and total lifecycle management are therefore not isolated regulatory parts. They are different segments supporting the same accelerated-access logic. Seeing only “can launch earlier” while missing the back-end monitoring, verification and patient-protection design misreads the nature of the system.

What this path means for the next decade

Part 1, from the angle of “data interfaces”, argued that complete public data is not the same as usable public data, and came with a four-language searchable tool page covering the 30 categories, their ICD-10 codes and validity periods. Part 2, from the angle of “reimbursement reception”, argued that the validity-period tiering of catastrophic illness already contains many of the elements required by AI-era tiered reimbursement. This installment, from the angle of “market entry”, shows how the dual Acts, conditional approval, and T-RMAT plug “Phase II evidence already mature” regenerative medicinal products into Taiwan’s launch and clinical-use pathway. Looking at the three together, Taiwan is no longer just discussing innovation in the abstract after the dual Acts came into force; an institutional framework for evaluating conditional access to specific products has begun to take concrete shape. For industry, the next real step is not another round of abstract discussion about “whether Taiwan matters”, but judging which product types, which evidence combinations, and which timing arrangements can actually move through this path.

For the global regenerative medicine industry, whether Taiwan is worth re-evaluating in the coming years will depend on whether this system can accumulate real-world success cases and whether the competent authorities can continue providing predictable review interaction. For the policy and review side, the completeness of this path will ultimately come back to two things: whether the safety record stays stable, and whether conditional approval can actually transition into full evidence completion downstream.

For patients and families, the real meaning of this design is time. The time gap of the same therapy launching in different countries has historically been the decisive factor for fast-progressing diseases. The core value of the dual Acts, conditional approval, and T-RMAT — taken together — is that this time gap is, in part, released earlier through the combination of post-market verification, risk management, and patient remedies. For some patients, this is not just policy progress; it is the possibility of “still being able to wait”.

For those working in cross-border collaboration, this pathway also offers a new starting point for conversation. Over the past decade or more, Taiwan-Japan collaboration has mostly clustered around electronics, semiconductors, and circular materials. After the dual Acts came into force, Taiwan and Japan now have an additional shared topic for discussion: whether Japan’s years of accumulated Sakigake and conditional-approval experience, and Taiwan’s newly launched conditional approval plus T-RMAT, can be linked into a workable cross-border path on actual cases. This is one of the directions this series will continue to track.

The next piece will move from the angle of “data extrapolation” — into ICH E5’s framework for bridging data, the definition of East Asian populations, and how, when global clinical data is being weighed for use in Taiwan, regulators and applicants typically think about population factors and external data acceptability. That will connect “market entry” in this piece to the actual technical detail of an application.

Glossary

  • Regenerative Medicine Act: Law governing the institutional execution of regenerative medicine technologies (clinical side). Promulgated June 2024, in force from 1 January 2026. See Part 1 glossary.
  • Regenerative Medicinal Products Act: Law governing the R&D, manufacturing, and marketing of regenerative medicinal products (product side, including cell therapy drugs). Promulgated June 2024, in force from 1 January 2026. Together with the Regenerative Medicine Act, often referred to as Taiwan’s “Regenerative Medicine Dual Acts”.
  • Conditional Approval (有附款許可): Under Article 9 of the Regenerative Medicinal Products Act. For regenerative medicinal products meeting the criteria of “life-threatening or seriously disabling disease”, “completed Phase II human trials”, and “data submitted by the applicant having been reviewed by the competent authority as showing safety and preliminary efficacy and cleared by the Regenerative Medicine Review Board”, conditional approval may be granted for up to five years, non-renewable.
  • T-RMAT (Taiwan Regenerative Medicine Advanced Therapy Pilot, 臺灣再生醫療製劑輔導專案): An accelerated guidance program promoted by TFDA and CDE, announced in March 2026. The focus is on pre-submission regulatory and scientific consultation, review preparation, and increased communication density. Officially stated target review timelines are 120 days for BLAs (including conditional approval) and 15 days for INDs; actual case outcomes still depend on data completeness. The naming logic mirrors the U.S. FDA’s RMAT designation.
  • Regenerative Medicine Review Board: A review body established under the Regenerative Medicinal Products Act, responsible for reviewing significant matters including conditional approval and efficacy verification plans.
  • FDA RMAT (Regenerative Medicine Advanced Therapy Designation): The U.S. FDA’s accelerated pathway for regenerative medicine, operating since the 21st Century Cures Act of 2016. For cell, tissue-engineering, and gene therapies that target serious or life-threatening diseases and show preliminary clinical evidence of addressing unmet medical needs, RMAT offers accelerated approval, rolling review, and intensified interaction.
  • EMA PRIME (PRIority MEdicines): The European Medicines Agency’s priority medicines scheme since March 2016, offering early CHMP/CAT rapporteur appointment, iterative scientific advice, and aligned accelerated assessment.
  • PMDA Sakigake: Japan’s MHLW “Sakigake (Pioneer)” designation since 2015, a forerunner accelerated review system applicable for comparison purposes. For products first developed in Japan with early data showing prominent effectiveness, Sakigake offers accelerated review; from 2017 onwards it has been further combined with the conditional approval pathway for regenerative medicinal products, forming Japan’s “accelerated design + earlier access” dual-layer architecture.
  • ICH E5: The International Council for Harmonisation guideline on ethnic factors and acceptability of foreign clinical data — the foundational framework for bridging data arguments. The next piece will go deeper into this.
  • PIC/S GMP: Pharmaceutical Inspection Co-operation Scheme — Good Manufacturing Practice.
  • GDP: Good Distribution Practice.
  • RWD (Real-World Data): Health-related data collected in routine medical settings (EMRs, NHI data, registries, patient-reported outcomes), usable to support efficacy verification during the conditional approval period.

Disclaimer: This article is a compilation of public information and an institutional observation. It does not constitute medical, legal, investment, clinical trial design, or pharmaceutical filing advice. Whether a specific product qualifies for conditional approval, whether it is suited to enter T-RMAT, and whether foreign clinical data can be accepted, all remain subject to the latest announcements of Taiwan’s TFDA and CDE, the Regenerative Medicine Act, the Regenerative Medicinal Products Act, and their related sub-regulations.

Sources: