---

Article 1 — Gorgulho J et al. — Circulating EVs carrying PD-1, PD-L1, CTLA-4 in HCC

PMID 42562419 | Gut | Prospective Cohort | Triage Score: 10

Dimension	Score	Rationale
Scientific Novelty	8	Multiplex EV-based immune checkpoint liquid biopsy is mechanistically novel; prior work exists on single checkpoint EVs but simultaneous PD-1/PD-L1/CTLA-4 in a validated prospective design is a meaningful advance
Clinical Relevance	9	HCC is a high-mortality cancer with ~70% ICI non-response; a 36–42 week lead time on imaging for acquired resistance is potentially transformative for clinical decision-making
Population Reach	7	HCC is the 6th most common cancer globally (~900,000 new cases/year); immunotherapy is now first-line, so the eligible population is large
Implementation Speed	6	EV isolation and multiplex checkpoint quantification are not yet CLIA/CE-IVD standardized; requires analytical validation and regulatory clearance, but the biomarker is blood-based
Evidence Strength	7	Prospective design with independent validation cohort is credible; abstract-only access and medium classification confidence cap this; sample size not confirmed

Key quantitative result: Predicts acquired resistance 36–42 weeks before radiographic detection; specific AUROCs not confirmed from abstract. External validation: Yes — training + independent validation cohort design reported. Main limitation: Abstract-only access; sample size unknown; EV isolation methodology variability across labs is a known reproducibility concern. Equity implications: HCC disproportionately affects men of Asian and African descent and patients with viral hepatitis — populations with limited access to advanced molecular testing. If this test becomes standard, equitable access pathways will be critical. Evidence Maturity: ✅ Confirmed — Validated (prospective + independent validation)

---

Article 2 — Kungwankiattichai S et al. — LBCL After CD19 CAR-T Failure

PMID 42562324 | Transplant Cell Ther | Multicenter Cohort | Triage Score: 9

Dimension	Score	Rationale
Scientific Novelty	6	Real-world characterization of post-CAR-T failure is important but not mechanistically novel; prognostic model adds modest novelty; value is primarily in updated data
Clinical Relevance	8	Post-CAR-T failure in LBCL has no established standard of care; prognostic models and real-world outcomes data directly guide treatment sequencing and trial enrollment
Population Reach	6	~18,000 LBCL cases/year in the US eligible for CAR-T; a meaningful but not enormous population; high unmet need within this group
Implementation Speed	7	Prognostic models from real-world data can be integrated into clinical practice relatively quickly without regulatory hurdles
Evidence Strength	6	Multicenter real-world design provides breadth; lacks randomization; abstract-only; sample size unknown; medium confidence

Key quantitative result: Not specified in abstract; characterizes prognostic factors and prediction models (specific HRs/OS rates not reported in triage data). External validation: Multicenter design provides partial external generalizability but explicit external validation cohort not confirmed. Main limitation: Observational design with inherent selection bias; no randomized comparison; treatment heterogeneity across sites. Equity implications: CAR-T access remains concentrated in academic medical centers; real-world data from multicenter study may inadvertently underrepresent community-treated patients and racial/ethnic minorities. Evidence Maturity: ✅ Confirmed — Validated (real-world, multicenter)

---

Article 3 — Lyu C et al. — Venetoclax Therapeutic Window Post-Transplant AML/MDS

PMID 42561008 | Hematol Oncol | Cohort Study | Triage Score: 9

Dimension	Score	Rationale
Scientific Novelty	7	Post-transplant venetoclax maintenance is an emerging and underexplored application; defining the therapeutic window (timing, dosing, safety) fills a real evidence gap with no dedicated Phase 3 data
Clinical Relevance	8	Post-transplant relapse in high-risk AML/MDS carries >80% mortality; dosing guidance for an approved agent addresses an immediate clinical need
Population Reach	6	~5,000–8,000 allogeneic SCTs/year in the US for AML/MDS; smaller population but very high unmet need
Implementation Speed	7	Venetoclax is already approved; dosing guidance from a cohort study can influence clinical practice relatively quickly, especially given lack of alternatives
Evidence Strength	5	Cohort design without randomization; abstract-only; design quality scored 1/3 by triage agent; no external validation confirmed; medium confidence

Key quantitative result: Defines optimal timing/dosing windows; specific OS/RFS outcomes not reported in triage data. External validation: Not confirmed; single-institution or limited-center likely. Main limitation: Non-randomized cohort; potential confounding by indication (sicker patients may not receive maintenance); small sample size likely. Equity implications: AML/MDS disproportionately affects older adults; access to allogeneic transplant and post-transplant monitoring is concentrated in well-resourced centers, creating access disparities. Evidence Maturity: ⬇️ Revised — Exploratory–Validated (cohort without randomization or external validation; "Validated" label from triage overstates rigor for this application)

---

Article 4 — Chen Z et al. — NEURAL-PC Deep Learning for NEPC Subtyping

PMID 42560769 | JCI Insight | External Validation Study | Triage Score: 9

Dimension	Score	Rationale
Scientific Novelty	8	H&E-only deep learning classifier for neuroendocrine prostate cancer using interpretable cellular features + MIL architecture is technically novel; replaces expensive molecular assays
Clinical Relevance	8	NEPC is an aggressive, rapidly lethal variant; correct subtyping changes treatment (platinum-based chemotherapy vs. AR-pathway inhibitors); H&E is universally available
Population Reach	6	NEPC comprises 17–25% of mCRPC (230,000 US mCRPC patients); meaningful but specialty population
Implementation Speed	5	FDA/CE digital pathology regulatory pathway needed; computational pathology infrastructure required at deploying institutions
Evidence Strength	8	External validation at major cancer centers, AUROC 0.921 is strong; full text available; multi-institutional validation design is the appropriate gold standard for AI diagnostic tools

Key quantitative result: AUROC 0.921 in independent external validation. External validation: ✅ Yes — explicitly multi-institutional external validation. Main limitation: Retrospective tissue samples; deployment requires digital pathology infrastructure; prospective clinical impact on treatment decisions not yet studied. Equity implications: Replacing expensive molecular assays with H&E-based AI could democratize NEPC diagnosis in resource-limited settings — a meaningful equity benefit if infrastructure barriers are addressed. Evidence Maturity: ✅ Confirmed — Validated (multi-institutional external validation)

---

Article 5 — van der Heijde D et al. — Deucravacitinib Phase 3 POETYK PsA-1

PMID 42562758 | Ann Rheum Dis | Phase 3 RCT | Triage Score: 8

Dimension	Score	Rationale
Scientific Novelty	7	First Phase 3 data for a TYK2-selective inhibitor in PsA; TYK2 selectivity over JAK1/2/3 is mechanistically and clinically differentiated
Clinical Relevance	8	PsA affects ~1M US patients; TYK2 selectivity avoids JAK inhibitor safety signals (thromboembolic risk, malignancy); 52-week structural damage data is clinically decisive
Population Reach	8	Psoriatic arthritis is common; patients who fail TNFi or IL-17i biologic classes represent a large secondary treatment population
Implementation Speed	6	Deucravacitinib is already approved for plaque psoriasis (Sotyktu); PsA indication approval likely near; regulatory pathway relatively clear
Evidence Strength	8	Phase 3 RCT, double-blind, placebo-controlled, 52-week follow-up; abstract-only limits full critical appraisal but design is high quality

Key quantitative result: ACR20 54.2% vs. 34.1% placebo at 16 weeks (P<0.001); sustained structural damage inhibition at 52 weeks. External validation: Inherent in Phase 3 RCT design; awaiting second Phase 3 trial (POETYK PsA-2) for full confirmation. Main limitation: Abstract-only; placebo-controlled (no active comparator arm vs. biologics); long-term safety beyond 52 weeks unknown; head-to-head vs. IL-17/IL-23 inhibitors not established. Equity implications: Oral availability is an equity advantage over injectable biologics; however, cost will likely be a significant barrier in lower-income settings. Evidence Maturity: ✅ Confirmed — Potentially Practice-Changing

---

Article 6 — Wakelee HA et al. — SKYSCRAPER-03 Tiragolumab + Atezolizumab vs. Durvalumab Stage III NSCLC

PMID 42562209 | Ann Oncol | Phase 3 RCT | Triage Score: 8

Dimension	Score	Rationale
Scientific Novelty	7	First Phase 3 head-to-head test of TIGIT blockade (tiragolumab) + PD-L1 vs. standard durvalumab in stage III NSCLC; negative result is itself definitive
Clinical Relevance	9	Definitively closes the question of tiragolumab+atezolizumab vs. durvalumab in unresectable stage III NSCLC; redirects clinical practice and trial design
Population Reach	8	Stage III NSCLC is the largest potentially curable lung cancer subgroup (~30,000 US patients/year); global burden is enormous
Implementation Speed	9	Immediate impact: negative trial result prevents adoption of an inferior/equivalent but more toxic/expensive regimen
Evidence Strength	8	Phase 3 RCT, randomized, head-to-head; HR 0.96 (95% CI 0.75–1.23) is definitive null; abstract-only caps full confidence slightly

Key quantitative result: PFS HR 0.96 (95% CI 0.75–1.23) — no improvement; OS data reportedly included. External validation: Phase 3 multicenter RCT is inherently externally validated. Main limitation: Abstract-only; subgroup analyses (by PD-L1 expression level, histology) not available for review; does not address whether tiragolumab fails across all NSCLC contexts. Equity implications: Negative result protects patients globally from exposure to a more complex, potentially more expensive regimen with no benefit; particularly relevant for health systems with limited oncology drug budgets. Evidence Maturity: ✅ Confirmed — Validated (definitive negative Phase 3)

---

Article 7 — Armenian SH et al. — Technology-Enabled Skin Cancer Screening After HCT

PMID 42562001 | JNCCN | RCT | Triage Score: 8

Dimension	Score	Rationale
Scientific Novelty	6	Technology-enabled patient activation is not novel in concept; novelty lies in application to post-HCT survivorship and the specific intervention design
Clinical Relevance	7	HCT survivors have 2–5× elevated skin cancer risk and this is a documented care gap; RCT evidence supports a scalable intervention
Population Reach	5	~25,000 allogeneic HCTs/year in the US; growing survivor population; relatively specialized
Implementation Speed	7	Technology-enabled interventions can be deployed at transplant centers without regulatory barriers; patient portal/app-based activation is feasible
Evidence Strength	7	RCT design; abstract-only limits appraisal of specific activation rates, skin cancer detection rates, and follow-through metrics

Key quantitative result: "Improved patient and physician activation" — specific rates/effect sizes not reported in triage data. External validation: Single-center RCT at City of Hope; generalizability to other transplant programs requires further study. Main limitation: Abstract-only; actual skin cancer detection rates (the clinically meaningful outcome) not confirmed; single-center design. Equity implications: HCT survivors include a high proportion of patients who survived aggressive hematologic malignancies; cancer surveillance is worse in lower-income and uninsured survivors. Technology-based tools may paradoxically worsen equity if digital access is not addressed. Evidence Maturity: ✅ Confirmed — Potentially Practice-Changing (for transplant centers with infrastructure)

---

Article 8 — Soler Espejo E et al. — AI Phenotyping of AF via GTM

PMID 42562389 | J Med Internet Res | Prospective Cohort | Triage Score: 7

Dimension	Score	Rationale
Scientific Novelty	7	GTM (generative topographic mapping) applied to AF phenotyping is methodologically novel; interpretable clustering distinguishes this from prior black-box AI approaches
Clinical Relevance	7	Four phenotypes with distinct 2-year risks of stroke, bleeding, and death could refine anticoagulation decisions beyond CHA₂DS₂-VASc alone
Population Reach	9	AF affects ~37 million people worldwide; even modest improvements in risk stratification have enormous aggregate impact
Implementation Speed	5	Requires clinical validation of phenotype-guided treatment decisions before adoption; regulatory pathway for AI risk stratification tools is complex
Evidence Strength	7	Large prospective cohort (n=3,259); 2-year follow-up; GYH Lip group's credibility; full text available; lacks randomized intervention arm

Key quantitative result: Four phenotypes with statistically distinct 2-year risks of thromboembolism, major bleeding, and all-cause death; specific C-statistics/event rates not reported in triage data. External validation: Not confirmed from triage data; single-cohort prospective. Main limitation: Phenotype-based risk stratification requires prospective validation that treatment decisions guided by phenotype improve outcomes vs. standard risk scores. Equity implications: AF is underdiagnosed in women and minority populations; phenotype-based tools could either reduce or amplify existing disparities depending on training data composition. Evidence Maturity: ✅ Confirmed — Validated (prospective large cohort; intervention validation still needed)

---

Article 9 — Yasuda Y et al. — Olaparib in Japanese BRCA-altered mCRPC

PMID 42559798 | Int J Urol | Cohort Study | Triage Score: 7

Dimension	Score	Rationale
Scientific Novelty	5	Confirmatory real-world data in a specific ethnic population; no mechanistic novelty; adds population-level generalizability
Clinical Relevance	6	Confirms olaparib efficacy in East Asian patients; clinically useful for Japanese oncologists; 14% BRCA PV prevalence finding has practical testing implications
Population Reach	4	Small cohort (n=27); Japanese BRCA-altered mCRPC is a narrow population; findings need confirmation
Implementation Speed	8	Olaparib already approved; this supports existing practice for a population previously underrepresented in trials
Evidence Strength	4	Very small n=27; abstract-only; single-institution; no control arm; medium confidence

Key quantitative result: Median OS 18.4 months; 1-year OS rate 80.6%; consistent with PROfound trial data. External validation: Consistency with PROfound provides indirect validation; not a new validation study. Main limitation: n=27 is underpowered; single-center; no subgroup analysis by BRCA1 vs. BRCA2; prior treatment lines vary. Equity implications: Directly addresses underrepresentation of East Asian patients in landmark PARP inhibitor trials — a meaningful equity contribution. Evidence Maturity: ⬇️ Revised — Exploratory (too small to be "Validated" for a new population; confirmatory but not definitive)

---

Article 10 — Ellard S et al. — BSGM Guidance on Incidental Findings in Rare Disease Genomics

PMID 42562627 | J Med Genet | Guidance Document | Triage Score: 7

Dimension	Score	Rationale
Scientific Novelty	5	Guidance documents codify existing evidence; the novelty is in formal policy articulation, not discovery
Clinical Relevance	8	Directly affects clinical workflow for every NHS genetics laboratory; standardizes a previously variable practice affecting patient communication and safety
Population Reach	7	Relevant to all patients undergoing genomic testing in the UK (hundreds of thousands/year); guidance will influence international programs
Implementation Speed	9	Published guidance from a national body can be implemented immediately in NHS laboratories
Evidence Strength	6	Guidance documents are not primary research; evidence synthesis quality depends on cited studies; full text available

Key quantitative result: N/A — policy document; 22 specific recommendations. External validation: N/A — expert consensus process. Main limitation: Guidance may not reflect the full range of international genomic testing contexts; evidence base for some recommendations may be limited. Equity implications: Clear guidance protects patients from harm of receiving un-counseled incidental variants; particularly important in resource-limited settings where genetic counseling access is constrained. Risk of inequity if guidance is UK-centric and not adapted globally. Evidence Maturity: ✅ Confirmed — Validated (as policy; not primary research)

---

Article 11 — Mathew AA et al. — CAR-T Strategies Targeting LSCs and MRD in AML (Review)

PMID 42561659 | Int Immunopharmacol | Review | Triage Score: 7

Dimension	Score	Rationale
Scientific Novelty	6	LSC-targeting CAR-T is a recognized frontier; a review synthesizing innovative approaches is useful but not itself novel
Clinical Relevance	7	AML CAR-T is urgently needed; review of strategies addressing the two main failure mechanisms (LSC persistence, MRD) is directly clinically relevant
Population Reach	7	AML affects ~20,000 new patients/year in the US; relapse is universal in high-risk groups
Implementation Speed	2	All strategies described are preclinical or early Phase 1; translation is years away
Evidence Strength	3	Review article; no primary data; abstract-only; limited quality appraisal possible

Key quantitative result: N/A — review. External validation: N/A — narrative review. Main limitation: Reviews can selectively emphasize promising findings; AML CAR-T has repeatedly failed in clinical trials despite promising preclinical data. Equity implications: AML outcomes are significantly worse in older adults and Black patients; effective CAR-T would have a substantial equity impact if made accessible. Evidence Maturity: ⬇️ Revised — Exploratory (review of early-stage strategies; "Validated" label from triage is inappropriate for a review in an early field)

---

Article 12 — Chen S et al. — AI-Driven PROTAC Design for CLIP1-LTK Fusion

PMID 42561017 | PNAS | Preclinical | Triage Score: 7

Dimension	Score	Rationale
Scientific Novelty	9	AI-guided PROTAC design for a specific oncogenic fusion protein is genuinely cutting-edge; combines two high-novelty fields (AI drug design + targeted protein degradation)
Clinical Relevance	3	Preclinical; CLIP1-LTK is an ultra-rare fusion; cannot exceed 5 per rules — scored 3 given early stage and rarity
Population Reach	2	CLIP1-LTK is extremely rare (identified in ~1–2% of lung adenocarcinomas); very small immediate population
Implementation Speed	1	Preclinical; first-in-human is likely 5–10+ years away
Evidence Strength	5	PNAS is high-quality; preclinical mixed model; abstract-only; cannot confirm rigor of in vivo studies

Key quantitative result: Overcomes kinase inhibitor resistance in preclinical model; specific tumor regression/IC50 data not reported in triage. External validation: None at this stage (preclinical proof-of-concept). Main limitation: Preclinical only; PROTAC pharmacokinetics and in vivo tolerability remain major hurdles; CLIP1-LTK rarity limits direct patient impact. Equity implications: Rare fusion oncogenes are being identified through comprehensive genomic profiling, which is not uniformly accessible; AI-designed precision therapeutics may widen rather than narrow treatment gaps if not paired with equitable genomic testing programs. Evidence Maturity: ✅ Confirmed — Exploratory

---

Article 13 — Subharam M et al. — cfDNA Methylation Blood Test for Early Cancer Detection in Military Personnel

PMID 42560195 | Mil Med | Cohort Study | Triage Score: 7

Dimension	Score	Rationale
Scientific Novelty	6	MCED liquid biopsy in a high-risk occupational cohort is a meaningful application but the underlying technology (cfDNA methylation + AI/ML) is established
Clinical Relevance	6	Proof-of-concept in a high-risk underserved population is clinically relevant; specific sensitivity/specificity data would elevate this score
Population Reach	6	~2.8 million active US military + large veteran population with documented elevated cancer risk from toxin exposure; broader relevance to all occupationally exposed workers
Implementation Speed	5	MCED technology is advancing toward clinical deployment; application to this population requires dedicated validation studies
Evidence Strength	4	Military Medicine supplement; abstract-only; small sample size suspected; cohort without prospective validation design; medium confidence

Key quantitative result: "Proof-of-concept" — specific sensitivity/specificity/detection rates not reported in triage data. External validation: Not confirmed. Main limitation: Supplement publication raises concerns about peer review rigor; sample size likely small; proof-of-concept framing limits conclusions. Equity implications: Directly targets an underserved high-risk population with documented gaps in cancer surveillance — a strong equity signal. Evidence Maturity: ⬇️ Revised — Exploratory (proof-of-concept; "Validated" label overstates the evidence from a supplement publication)

---

Article 14 — Koskan AM et al. — Implementation Strategies for Cardiac Amyloidosis Early Detection

PMID 42558576 | Front Cardiovasc Med | Qualitative/Pre-implementation Study | Triage Score: 7

Dimension	Score	Rationale
Scientific Novelty	5	Implementation science applied to ATTR-CA detection is valuable but the strategies themselves (AI-EHR integration, scintigraphy expansion) are not novel in isolation
Clinical Relevance	7	ATTR-CA is significantly underdiagnosed; tafamidis is effective when started early; implementation barriers are well-documented and this framework is actionable
Population Reach	6	ATTR-CA prevalence estimated at ~13% of HFpEF patients; affects millions globally if properly screened
Implementation Speed	7	Qualitative frameworks can be adopted immediately by health systems; no regulatory barriers to implementing the recommended strategies
Evidence Strength	4	Qualitative pre-implementation study; no outcomes data; context is VA hospital system which may limit generalizability

Key quantitative result: 23 implementation strategies identified across multiple domains. External validation: N/A — qualitative study. Main limitation: Pre-implementation; no outcomes data showing that implementation of these strategies improves ATTR-CA detection rates; VA context may limit generalizability. Equity implications: ATTR-CA disproportionately affects Black men (Val122Ile variant) who are systematically underdiagnosed; implementation strategies specifically addressing this disparity would be high-value. Evidence Maturity: ⬇️ Revised — Exploratory (pre-implementation; "Validated" is inappropriate without outcomes data)

---

Article 15 — Mazzola A et al. — GLP-1 RAs and SGLT-2i in Liver Transplant Recipients

PMID 42562072 | Clin Res Hepatol Gastroenterol | Cohort Study | Triage Score: 7

Dimension	Score	Rationale
Scientific Novelty	7	GLP-1/SGLT2i use in liver transplant recipients is novel; this population was excluded from landmark trials; 80% insulin withdrawal is a striking finding
Clinical Relevance	7	Post-transplant diabetes affects ~20–40% of recipients; insulin management is burdensome and has rejection/infection implications; this evidence fills a real practice gap
Population Reach	4	~9,000 liver transplants/year in the US; relatively small but high-need population
Implementation Speed	7	GLP-1 RAs and SGLT2i are commercially available; transplant hepatologists can apply findings to current patients with appropriate caution
Evidence Strength	4	n=41; single-center; cohort; abstract-only; medium confidence — small sample size is the dominant limitation

Key quantitative result: 90% HbA1c target achievement; 80% insulin withdrawal in new-onset post-transplant diabetes. External validation: Not confirmed. Main limitation: n=41 is very small; potential confounding by graft function, immunosuppression type; drug interaction risks with calcineurin inhibitors warrant careful assessment. Equity implications: Liver transplant recipients are already a health-system-engaged population with access to specialist care; equity impact is modest but important for improving metabolic outcomes in this group. Evidence Maturity: ⬇️ Revised — Exploratory (too small and uncontrolled to be "Validated")

---

Article 16 — Jeemon P et al. — Family-Based Lipid Interventions Cluster RCT India

PMID 42562562 | J Clin Lipidol | Cluster RCT | Triage Score: 6

Dimension	Score	Rationale
Scientific Novelty	4	Family-based CVD interventions are well-established in the literature; LMIC context adds some novelty
Clinical Relevance	5	Relevant to primary prevention; effect size and which lipid parameters improved not specified in triage data
Population Reach	8	India has the world's second-largest CVD burden; scalable family-based interventions could affect tens of millions
Implementation Speed	6	Low-technology intervention potentially scalable in primary care settings without regulatory barriers
Evidence Strength	6	Cluster RCT is methodologically appropriate for this type of intervention; abstract-only limits full appraisal

Key quantitative result: "Improved lipid outcomes" — specific LDL-C reductions or lipid targets not reported in triage data. External validation: Not confirmed. Main limitation: Abstract-only; effect size unknown; cluster contamination is a risk; long-term CVD event outcomes not addressed. Equity implications: LMIC-focused; directly targets populations with least access to pharmacotherapy; a meaningful equity focus. Evidence Maturity: ✅ Confirmed — Validated (cluster RCT design, though effect size and durability remain uncertain)

---

Article 17 — Kyryk V et al. — Cellular Senescence Markers in Post-COVID Cardiovascular Patients

PMID 42557844 | Biomed Res Int | Cohort Study | Triage Score: 6

Dimension	Score	Rationale
Scientific Novelty	5	Post-COVID senescence is an active area; cytological markers in CV patients adds specificity but is not groundbreaking
Clinical Relevance	4	Mechanistic/descriptive; no therapeutic intervention; not yet actionable for patient care
Population Reach	7	Global COVID survivor population is enormous; CV complications are a major long-COVID burden
Implementation Speed	3	No therapeutic implication yet; senolytic trials are ongoing but this study doesn't advance them directly
Evidence Strength	4	Biomed Res Int is lower-tier; cohort design; sample size unknown; full text available but medium confidence

Key quantitative result: Elevated cytological senescence markers in post-COVID CV patients vs. controls — specific marker levels not reported in triage data. External validation: Not confirmed. Main limitation: Descriptive; causality not established; lower-impact journal; unknown sample size. Equity implications: Post-COVID health impacts are disproportionate in lower-income and minority populations; mechanistic understanding could eventually support targeted interventions for these groups. Evidence Maturity: ⬇️ Revised — Exploratory (descriptive biomarker study; "Validated" label overstates this)

---

Article 18 — Rapino C et al. — Steatosis and Alcohol Impact on Mortality (Marginal Structural Models)

PMID 42557925 | Liver Int | Cohort Study | Triage Score: 6

Dimension	Score	Rationale
Scientific Novelty	6	Marginal structural models applied to time-varying liver disease exposures in a well-characterized longitudinal cohort is methodologically sophisticated
Clinical Relevance	5	Causal quantification of steatosis + alcohol on mortality is informative for counseling; not immediately practice-changing
Population Reach	7	Fatty liver disease and alcohol use affect tens of millions globally; population-level relevance is high
Implementation Speed	4	Findings inform risk communication rather than immediate treatment changes
Evidence Strength	6	Liver Int is credible; marginal structural models are methodologically rigorous; Canadian Longitudinal Study on Aging is well-characterized; abstract-only

Key quantitative result: Time-varying causal estimates for steatosis and alcohol on all-cause mortality — specific HR values not reported in triage data. External validation: Not confirmed. Main limitation: Observational; residual confounding possible even with MSMs; steatosis likely assessed by FIB-4 or imaging rather than biopsy. Equity implications: Alcohol-related liver disease disproportionately affects socioeconomically disadvantaged populations; better causal understanding can support policy interventions. Evidence Maturity: ✅ Confirmed — Validated (methodologically rigorous observational study)

---

Article 19 — Zhang H et al. — miR-126 Prognostic Value in AML

PMID 42562619 | Int J Lab Hematol | Cohort Study | Triage Score: 5

Dimension	Score	Rationale
Scientific Novelty	3	miR-126 in AML is well-established in the literature; limited additional novelty
Clinical Relevance	4	Prognostic biomarker data; not immediately actionable without therapeutic implications
Population Reach	6	AML affects ~20,000/year in the US; biomarker findings could eventually influence risk stratification
Implementation Speed	4	Single miRNA biomarkers rarely enter clinical practice without a validated treatment pathway
Evidence Strength	4	Cohort; abstract-only; conservative per triage agent's own scoring; unknown sample size; medium confidence

Key quantitative result: miR-126 has prognostic value — specific OS/RFS HR not reported. External validation: Not confirmed. Main limitation: Abstract-only; single miRNA biomarker with modest incremental value over existing cytogenetic/molecular risk stratification. Equity implications: Minimal specific equity considerations; routine miRNA testing is not widely deployed. Evidence Maturity: ⬇️ Revised — Exploratory (single-institution, abstract-only, limited novelty)

---

Article 20 — Khosravifarsani M et al. — Serum Biomarkers and CLL Stage

PMID 42561705 | Hematol Transfus Cell Ther | Cohort Study | Triage Score: 5

Dimension	Score	Rationale
Scientific Novelty	3	Biomarker-stage correlations in CLL are extensively studied; limited new ground broken
Clinical Relevance	3	Descriptive associations without prospective validation or therapeutic implications
Population Reach	6	CLL is the most common leukemia in adults in Western countries
Implementation Speed	4	Not immediately implementable without prospective validation
Evidence Strength	3	Abstract-only; conservative score per triage policy; descriptive cohort; unknown size

Key quantitative result: Correlation coefficients between serum biomarkers and Binet/Rai staging — specific values not reported. External validation: Not confirmed. Main limitation: Purely descriptive; does not advance treatment algorithms. Equity implications: Minimal specific equity considerations. Evidence Maturity: ⬇️ Revised — Exploratory

---

Conflict Summary

No major direct conflicts exist across articles. One notable scientific tension: the SKYSCRAPER-03 negative result (Article 6) contradicts earlier optimism about TIGIT blockade, which aligns with a pattern of negative anti-TIGIT Phase 3 results accumulating in the field (Merck's vibostolimab, Roche's tiragolumab). This is a consistent signal, not a conflict within this batch. Articles 1 (EV biomarker, HCC) and 8 (AI phenotyping, AF) are complementary in demonstrating that AI/biomarker-enabled precision risk stratification is maturing across organ systems.

---

Ranked Table

Rank	Article	Flag	Impact Score	Clinical Relevance (30%)	Population Reach (25%)	Scientific Novelty (20%)	Implementation Speed (15%)	Evidence Strength (10%)	Triage Score	Study Design	Why It Matters
1	6 — SKYSCRAPER-03 NSCLC	⬜	8.05	9	8	7	9	8	8	Phase 3 RCT	A definitive Phase 3 negative trial resolves a major clinical question immediately, preventing adoption of an inferior regimen across ~30,000 US patients/year with stage III NSCLC. Negative results of this quality are practice-shaping.
2	1 — EV Checkpoint Biomarkers HCC	⬜	7.80	9	7	8	6	7	10	Prospective Cohort	A 36–42 week lead time over imaging for predicting ICI resistance in HCC is clinically extraordinary if validated; addresses ~70% of patients who fail immunotherapy in one of the fastest-growing cancer types globally.
3	5 — Deucravacitinib Phase 3 PsA	⬜	7.60	8	8	7	6	8	8	Phase 3 RCT	Phase 3 confirmation of the first-in-class TYK2-selective oral agent in PsA with a favorable safety profile and structural damage prevention through 52 weeks; regulatory approval likely near and could reshape treatment sequencing for ~1M US PsA patients.
4	4 — NEURAL-PC DL NEPC Subtyping	⬜	7.20	8	6	8	5	8	9	Validation Study	External validation at AUROC 0.921 for an H&E-based AI classifier replacing expensive molecular assays for neuroendocrine prostate cancer; could democratize diagnosis in resource-limited settings and accelerate time-to-treatment for a rapidly lethal variant.
5	8 — AI AF Phenotyping GTM	⬜	7.10	7	9	7	5	7	7	Prospective Cohort	Interpretable AI phenotyping of AF in 3,259 prospectively followed patients identifies four clinically distinct risk groups; AF's enormous global burden means even modest refinements in anticoagulation risk stratification translate to thousands of prevented strokes annually.
6	2 — LBCL Post-CAR-T Failure	⬜	6.80	8	6	6	7	6	9	Multicenter Cohort	Multicenter real-world data and prediction models for the highest-unmet-need space in aggressive lymphoma; directly informs salvage treatment sequencing and trial enrollment in an area where no standard of care exists post-CAR-T failure.
7	3 — Venetoclax Post-Transplant AML/MDS	⬜	6.80	8	6	7	7	5	9	Cohort Study	Defines the practical dosing and safety window for an approved agent in a post-transplant setting where post-transplant relapse is near-universally fatal; fills a critical evidence gap for hematologists already using venetoclax off-label in this context.
8	10 — BSGM Genomic Incidental Findings Guidance	⬜	6.75	8	7	5	9	6	7	Guidance Document	National-level guidance directly standardizing clinical genomics practice for all rare disease testing programs in the UK; immediately implementable and likely to influence international standards; protects patients from harm of un-counseled incidental findings.
9	7 — Technology-Enabled Skin Cancer Screening Post-HCT	🔴	6.55	7	5	6	7	7	8	RCT	RCT evidence for a scalable technology intervention closing a documented care gap in HCT survivors with 2–5× elevated skin cancer risk; directly applicable to transplant programs with existing patient portal infrastructure.
10	15 — GLP-1/SGLT2i in Liver Transplant Recipients	🟡	6.05	7	4	7	7	4	7	Cohort Study	First evidence supporting safety and efficacy of GLP-1/SGLT2i in liver transplant recipients with post-transplant diabetes; 80% insulin withdrawal rate is striking and immediately relevant to transplant hepatologists managing this common complication.
11	14 — Cardiac Amyloidosis Implementation Strategies	⬜	5.90	7	6	5	7	4	7	Qualitative Study	Twenty-three actionable implementation strategies for ATTR-CA early detection, where treatment exists but diagnosis is systematically delayed; VA hospital context provides a real-world roadmap for health system implementation.
12	13 — cfDNA Methylation Blood Test Military Personnel	🟡	5.75	6	6	6	5	4	7	Cohort Study	Proof-of-concept application of MCED liquid biopsy to a high-risk, underserved occupational cohort; demonstrates feasibility in a population with documented elevated cancer risk from toxin exposure and historically poor cancer surveillance.
13	12 — AI PROTAC Design CLIP1-LTK	⬜	4.50	3	2	9	1	5	7	Preclinical	Exceptional scientific novelty — AI-guided PROTAC design in PNAS for a resistance-prone oncogenic fusion — but preclinical stage and extreme rarity of target fusion protein limit near-term clinical impact; a landmark methodology paper worth tracking.
14	16 — Family-Based Lipid Interventions India Cluster RCT	🟡	5.65	5	8	4	6	6	6	Cluster RCT	Solid cluster RCT design addressing the enormous Indian CVD burden with a scalable, low-technology family-based intervention; population reach is the primary strength, though effect size and novelty are modest.
15	11 — CAR-T LSC/MRD Strategies in AML (Review)	⬜	5.35	7	7	6	2	3	7	Review	Clinically important synthesis of an urgent area (AML CAR-T has lagged far behind B-cell malignancy CAR-T); LSC persistence is the recognized central obstacle; low evidence strength (review) and distant implementation timeline limit score.
16	18 — Steatosis and Alcohol on Mortality (MSM)	⬜	5.30	5	7	6	4	6	6	Cohort Study	Methodologically sophisticated causal analysis of modifiable liver disease risk factors in a large longitudinal cohort; primarily informs population risk communication rather than clinical decision-making in individuals.
17	9 — Olaparib Real-World Japanese mCRPC	🟡	5.25	6	4	5	8	4	7	Cohort Study	Provides important real-world confirmation of olaparib efficacy in an ethnically underrepresented population; findings support existing practice and are immediately applicable to Japanese oncologists; sample size severely limits confidence.
18	8 — Post-COVID Senescence Markers CV Patients	⬜	4.50	4	7	5	3	4	6	Cohort Study	Documents elevated cellular senescence markers in post-COVID CV patients, adding to the mechanistic understanding of long-COVID cardiovascular burden; descriptive and not yet actionable; lower-impact journal limits confidence.
19	19 — miR-126 Prognostic Value AML	⬜	4.15	4	6	3	4	4	5	Cohort Study	Confirmatory prognostic biomarker data on a known AML-associated miRNA; limited novelty and unclear incremental clinical utility over established molecular risk stratification; primarily of interest as a component of future multi-marker panels.
20	20 — Serum Biomarkers and CLL Stage	⬜	3.65	3	6	3	4	3	5	Cohort Study	Descriptive biomarker-staging correlation in treatment-naive CLL; no prospective validation or therapeutic implications; limited novelty in a heavily studied area; lowest priority in this batch.

---

Impact Score formula: (Clinical Relevance × 0.30) + (Population Reach × 0.25) + (Scientific Novelty × 0.20) + (Implementation Speed × 0.15) + (Evidence Strength × 0.10)

---

PHASE 4 — Deep Dives

---

---

[HOOK]

Liver cancer kills nearly a million people every year, and immunotherapy was supposed to change that. And it has — for some. But roughly seven out of ten patients with hepatocellular carcinoma don't respond to immune checkpoint inhibitors, and by the time a scan shows the cancer progressing, months of ineffective treatment may have passed. What if a simple blood test could warn you — nearly a year in advance — that a patient is about to become resistant?

[THE DISCOVERY]

Researchers led by Johann von Felden and colleagues, publishing in Gut, found that tiny particles shed into the bloodstream — called extracellular vesicles, or EVs — carry immune checkpoint proteins on their surface. Specifically, they measured EVs carrying PD-1, PD-L1, and CTLA-4, the same three checkpoint molecules targeted by modern immunotherapy drugs. In a prospective study with an independent validation cohort, the combination of these EV-carried checkpoints distinguished patients who would respond to anti-PD-L1-based therapy from those who wouldn't. Even more striking: the EV signature predicted when a patient was about to develop acquired resistance — approximately 36 to 42 weeks before conventional CT or MRI scanning could detect the cancer progressing.

Think of it like a weather warning system. By the time you see the storm on the radar, it's already there. This blood test catches the atmospheric pressure dropping weeks before the clouds form.

[THE SCIENCE BEHIND IT]

This was a prospective cohort study — researchers enrolled patients before the outcome was known and tracked them forward in time, which is a stronger design than looking backwards at historical records. Critically, the findings were replicated in an independent validation cohort, the gold standard for a biomarker study of this type. The EV analysis measured surface expression of three checkpoint molecules simultaneously, allowing a multiplex signal that is more informative than any single marker alone.

The major limitation is that we're working from abstract-only access — the sample size, specific AUROCs, and the exact thresholds used for the EV signature have not been confirmed from the published text. EV isolation methodology also varies significantly between labs, which creates a real-world reproducibility challenge that will need to be addressed before clinical deployment.

[WHO THIS HELPS]

This is most immediately relevant for the estimated 900,000 patients diagnosed with hepatocellular carcinoma globally each year — the majority of whom now receive immunotherapy as first-line treatment. It is especially important for patients whose tumors cannot be safely biopsied, which is common in HCC. The benefit extends to clinicians who are currently flying blind between imaging timepoints, particularly in under-resourced settings where imaging frequency is constrained.

[THE REAL-WORLD IMPACT]

If this test enters clinical practice, the workflow changes profoundly. Instead of waiting for a scan to confirm what a blood marker already predicted nine months ago, oncologists could switch treatment strategies earlier — adding on CTLA-4 blockade, enrolling patients in trials of combination therapy, or transitioning to locoregional treatments before the cancer has spread further. Earlier detection of resistance could also reduce the number of patients who receive futile immunotherapy cycles, lowering costs and toxicity exposure.

[WHAT WE STILL DON'T KNOW]

The central unanswered question is whether acting on this early signal — switching therapy when the EV test turns positive — actually improves survival. Predicting resistance 36 weeks early is remarkable, but only matters clinically if there is an effective alternative therapy ready to deploy and if early switching proves superior to waiting for radiographic confirmation. That requires prospective interventional trials, which do not yet exist. Standardization of EV measurement across clinical labs is also an unsolved problem.

[LIKELIHOOD OF MAKING A DIFFERENCE]

• Scientific Confidence: High — prospective design with independent validation in a credible journal
• Translation Speed: 2–5 years for biomarker-only adoption; 5–10 years for treatment-decision validation
• Barrier Analysis:
  • Regulatory: Will require FDA/EMA IVD clearance as a companion diagnostic
  • Reimbursement: Payer acceptance will depend on interventional trial data showing clinical utility
  • Infrastructure: EV isolation and multiplex quantification are not standard in most clinical labs; analytical standardization is needed
  • Equity: HCC disproportionately affects patients in Southeast Asia and Sub-Saharan Africa — populations with the least access to advanced molecular diagnostics. Equitable access must be designed in from the start.

[CALL TO ACTION / CLOSING]

A blood marker that reads the future of immunotherapy resistance in liver cancer — nine months before a scan can see it — is exactly the kind of tool that could shift oncology from reactive to truly predictive. If the interventional evidence catches up to this discovery, patients won't just know sooner. They'll have time to act.

---

---

[HOOK]

CAR-T cell therapy was one of oncology's most celebrated breakthroughs for aggressive lymphoma. But for a significant proportion of patients, the celebration ends too soon. When a large B-cell lymphoma progresses after CD19-targeted CAR-T, patients and their oncologists enter medically uncharted territory — no approved standard of care, no clear consensus on what comes next, and outcomes that are often devastating. Understanding who survives this scenario, and why, is among the most urgent questions in hematologic oncology right now.

[THE DISCOVERY]

Researchers from the Oregon Health & Science University and collaborating centers, publishing in Transplantation and Cellular Therapy, analyzed updated multicenter real-world data from patients with large B-cell lymphoma who progressed after CD19 CAR-T cell therapy. Their work characterizes actual outcomes in this population — something clinical trials can't easily do because patients who fail CAR-T are rarely enrolled in subsequent prospective studies — and goes further by identifying specific prognostic factors and building prediction models that clinicians can use at the bedside.

[THE SCIENCE BEHIND IT]

This is a real-world multicenter cohort study, which means the data comes from actual clinical practice across multiple institutions rather than a controlled trial. That design has both strengths and weaknesses. The strength is that it captures the true heterogeneity of patients who receive CAR-T in routine care — including older patients, those with prior organ dysfunction, and those who received CAR-T later in their treatment course. The weakness is the absence of randomization, meaning that the prognostic factors identified could reflect who got treated with what, rather than pure biology. Specific survival outcomes and the exact factors in the prediction model were not reported in the triage data, as this is abstract-only access.

The study is built on multicenter data, which improves the reliability of the prognostic model compared to single-institution series. However, external prospective validation of the prediction model itself — testing it on new patients at new centers — will be needed before it should be used to guide therapy decisions.

[WHO THIS HELPS]

This directly impacts patients with relapsed or refractory large B-cell lymphoma who have already failed at least one line of therapy plus CD19 CAR-T — a group estimated at several thousand patients per year in the United States alone, with numbers growing as CAR-T use expands. It also helps oncologists at community cancer centers who increasingly manage CAR-T-treated patients but have limited access to expert LBCL consultation.

[THE REAL-WORLD IMPACT]

The prognostic model has two immediate applications. First, patient counseling: oncologists can more accurately describe prognosis to patients and families, enabling more realistic goal-setting and earlier palliative care conversations where appropriate. Second, clinical trial enrichment: identifying patients most likely to benefit from experimental salvage regimens — bispecific antibodies, novel CAR-T constructs, antibody-drug conjugates — can improve trial design and accelerate regulatory approval for this population. Without this kind of real-world evidence base, oncologists are making sequential therapy decisions in an evidence vacuum.

[WHAT WE STILL DON'T KNOW]

The critical unknown is whether the prediction model actually improves outcomes when used prospectively to guide therapy selection. A prognostic score that tells you who will do poorly does not automatically tell you what to do differently for those patients. It remains unclear whether any of the available post-CAR-T salvage options — including bispecific T-cell engagers, loncastuximab, or a second CAR-T product — preferentially benefit patients in specific prognostic groups. That question requires prospective biomarker-enriched trials.

[LIKELIHOOD OF MAKING A DIFFERENCE]

• Scientific Confidence: Moderate — multicenter real-world design is credible but observational; model validation pending
• Translation Speed: 2–5 years for prognostic model adoption; therapeutic implications depend on parallel trial progress
• Barrier Analysis:
  • Regulatory: Prognostic tools don't require regulatory approval; therapeutic decisions are the next hurdle
  • Reimbursement: Salvage therapies post-CAR-T are expensive and payer coverage is inconsistent
  • Infrastructure: CAR-T programs are concentrated in academic centers; prognostic model deployment requires digital clinical decision support tools
  • Equity: Access to CAR-T itself is already unequal; patients who fail CAR-T at community referral centers may not have access to academic programs for salvage therapy

[CALL TO ACTION / CLOSING]

When CAR-T fails in aggressive lymphoma, patients deserve more than uncertainty — they deserve a roadmap. This study is the beginning of one, and every future trial in the post-CAR-T space should be built on exactly this kind of real-world foundation.

---

---

[HOOK]

For patients with high-risk acute myeloid leukemia or myelodysplastic syndromes who make it through the grueling process of an allogeneic stem cell transplant, the next fear is relapse. Post-transplant relapse in AML and MDS is among the most feared outcomes in hematology — survival is typically measured in months, and treatment options are severely constrained by the fact that the patient just received a full conditioning regimen. Venetoclax, a BCL-2 inhibitor that transformed AML therapy, is increasingly being used after transplant to try to prevent relapse. But how do you use it safely? Too early, and it may harm the new graft. Too much, and it triggers life-threatening infections. Until now, no one has formally defined the boundaries.

[THE DISCOVERY]

Lyu C and colleagues, publishing in Hematological Oncology, systematically characterized the optimal timing, dosing, and safety limits for venetoclax when used as maintenance therapy after allogeneic stem cell transplant in high-risk AML and MDS patients. This is what physicians mean by a "therapeutic window" — the dosing space where the drug does enough good to justify the risk, without pushing into toxicity that harms the newly engrafted immune system or suppresses healthy hematopoiesis.

[THE SCIENCE BEHIND IT]

This is a cohort study — patients who received post-transplant venetoclax at a center experienced in this approach were followed prospectively and their outcomes, toxicities, and treatment parameters recorded and analyzed. The study's strength is that it directly addresses a gap that no Phase 3 trial has filled: the specific question of post-transplant venetoclax use. In the real world, hematologists have been extrapolating from pre-transplant AML trials to make these decisions, which is an imperfect approach given the completely different physiological context post-engraftment.

The limitation is significant: without randomization, we cannot be certain whether the observed benefit is due to venetoclax itself or to patient selection — it's possible that only patients with favorable graft kinetics and stable engraftment were offered maintenance, which would make the outcomes look better than they truly are for an unselected population. The abstract-only access means specific numbers on relapse-free survival, overall survival, and toxicity rates cannot be independently confirmed.

[WHO THIS HELPS]

This is directly relevant for an estimated 5,000–8,000 patients per year in the United States who undergo allogeneic transplant for AML or MDS, and more specifically for the substantial proportion who are classified as high-risk — those with adverse-risk cytogenetics, TP53 mutations, or complex karyotypes who are most likely to relapse after transplant. It also helps the transplant hematologists who care for these patients and currently lack any formalized dosing guidance for post-transplant venetoclax.

[THE REAL-WORLD IMPACT]

Defining a therapeutic window has immediate practical consequences. It tells transplant physicians when to start venetoclax after transplant (the timing question is critical — too early may interfere with engraftment), what dose to use in the post-transplant context where marrow reserve is limited, and what toxicity signals should prompt dose reduction or discontinuation. This kind of practical dosing guidance — even from a non-randomized study — fills the knowledge gap that clinicians face every day when they reach for venetoclax off-label in the post-transplant setting.

[WHAT WE STILL DON'T KNOW]

The fundamental unknown is whether post-transplant venetoclax maintenance actually reduces relapse rates and prolongs survival compared to observation, in a prospectively randomized comparison. Multiple randomized Phase 3 trials of post-transplant maintenance in AML (including azacitidine-based approaches) have previously failed to show benefit despite promising Phase 2 data. Venetoclax may be different — its mechanism directly targets BCL-2-dependent MRD clones — but that must be proven rigorously. Additionally, the interaction between venetoclax and graft-versus-host disease prophylaxis regimens needs careful characterization.

[LIKELIHOOD OF MAKING A DIFFERENCE]

• Scientific Confidence: Moderate — fills a real evidence gap but cohort study design limits causal inference
• Translation Speed: 2–5 years for guideline incorporation pending randomized data; immediate influence on off-label use is likely
• Barrier Analysis:
  • Regulatory: Venetoclax is already approved; post-transplant use is off-label; dosing guidance from cohort data is insufficient for label expansion without RCT
  • Reimbursement: Post-transplant maintenance is increasingly covered when data supports it; a cohort study may not satisfy payer requirements
  • Infrastructure: Requires transplant programs with experienced monitoring protocols; not available in all centers
  • Equity: Access to allogeneic transplant already reflects significant socioeconomic and geographic disparities; post-transplant maintenance options further concentrate specialized care in academic settings

[CALL TO ACTION / CLOSING]

Venetoclax reshaped AML therapy before transplant — this study begins to define whether it can protect the patients who made it to the other side. The therapeutic window has been sketched; now it needs to be drawn in ink with a randomized trial.