JNCI Journal of the National Cancer Institute最新文献

Background: Large-scale randomized controlled trials (RCTs) have established compelling evidence that screening by flexible sigmoidoscopy reduces colorectal cancer (CRC) incidence. Reported incidence results include cancers that were already prevalent and yet undiagnosed, but no longer preventable at screening. We aimed to derive, disentangle, and fully disclose early-detection and long-term prevention effects of screening sigmoidoscopy from published trial results.

Methods: We used data from 3 large-scale RCTs from the United Kingdom (UKFSST), Italy (SCORE), and the United States (PLCO), which included a total number of 359 198 participants. For each trial and each length of follow-up, we derived the numbers and proportions of CRC cases that were either early detected or prevented among screening attenders.

Results: In the UKFSST, which reported the longest follow-up data, screening sigmoidoscopy prevented 64% (95% CI = 59% to 69%) of incident distal CRC that would have been expected in the absence of screening during a median of 21.3 years. Within follow-up periods between 10 and 12 years, the proportions of distal CRC cases that were either early detected or prevented among screening users ranged between 67% (95% CI = 61% to 72%) in the PLCO and 80% (95% CI = 68% to 89%) in the SCORE trial, with approximately equal shares of early-detected and prevented cases in the SCORE and the PLCO trials, and a higher share of prevented cases in the UKFSST.

Conclusions: A single screening sigmoidoscopy prevents 2 out of 3 incident cancers in the distal colon and rectum over a period of more than 20 years, on top of early-detecting prevalent cases at screening.

Background: There is growing interest in artificial intelligence (AI) models for predicting future breast cancer. We performed a systematic review of studies of mammography-based AI models for future breast cancer risk prediction to summarize current evidence, identify knowledge gaps, and inform future research directions.

Methods: We searched 6 databases for studies from January 1, 2012, to February 28, 2025, that evaluated mammography-based AI models for future breast cancer risk prediction. We extracted study design, participants' race and ethnicity, geographic origin, mammogram type, vendor, prediction time frame, breast cancer type predicted, external validation, and exclusion of cancers diagnosed on the index screening mammogram. Areas under the receiver operating curve (AUCs) were summarized overall and by study characteristics.

Results: A total of 41 studies met our inclusion criteria. All studies were retrospective, and most used 2D mammograms (n = 37 studies) acquired using Hologic equipment (n = 25) and performed in the United States (n = 17); White, non-Hispanic women were most represented. Nearly all (n = 40) studies assessed discrimination performance with a median AUC of 0.71 for no longer than 2-year risk prediction, 0.72 for 3-4 years, and 0.71 for 5 years or more prediction. Median AUC was 0.75 for studies including index cancers vs 0.68 when excluded. Six studies reported model calibration performance ranging from good to overestimation of risk.

Conclusion: Future studies should evaluate models using digital breast tomosynthesis, examine performance for aggressive or advanced breast cancer, include diverse populations, and evaluate both discrimination and model calibration. Prospective evaluations are needed to determine the clinical utility of mammography-based AI models for personalized risk-based breast cancer screening before implementation.

Cannabis use for symptom management among patients with cancer has increased significantly in recent years, with many reporting benefits for pain, anxiety, sleep, nausea, appetite as well as other symptoms. However, rigorous prospective data on the potential benefits and harms of cannabis use in this population are lacking. This Commentary describes a United States National Cancer Institute (NCI)-led initiative addressing this research gap by supporting five prospective observational studies evaluating the benefits and harms of cannabis use among a large, heterogeneous samples of patients with cancer undergoing active systemic treatment. We provide an overview of each study, including cancer type, treatment modalities, inclusion/exclusion criteria, data collection methods, and both patient-reported and cancer-related outcomes.

Background: Despite well-documented national declines in cervical cancer incidence among young women following human papillomavirus (HPV) vaccine implementation, state-level data remain limited.

Methods: Using the US Cancer Statistics Database, differences in cervical cancer incidence rates for women aged 20-31 between the pre-vaccination (2000-2005) and vaccination era (2016-2021) were estimated using rate ratios (RRs) across 47 states and the District of Columbia (DC). Associations between HPV vaccination rates from the National Immunization Survey-Teen and RRs were examined using Spearman's rank test and linear regression models, adjusted for screening rates from the Behavioral Risk Factor Surveillance System.

Results: Nationwide, cervical cancer incidence rates declined by 27% (RR = 0.73, 95%CI:0.70 to 0.75) during the vaccination era, from 5.1 to 3.7 per 100,000. Reductions exceeded 50% in DC (RR = 0.48, 95%CI : 0.15 to 0.81), Rhode Island (RR = 0.48, 95%CI : 0.21 to 0.76), Michigan (RR = 0.48, 95%CI : 0.38 to 0.57), and Hawaii (RR = 0.49, 95%CI : 0.21 to 0.78), with 28 additional states achieving statistically significant reductions of 15-50%. Ten states showed slower decreases (<15%). Notably, progress was lacking in Vermont (RR = 1.11; 95%CI : 0.21 to 2.00), West Virginia (RR = 1.09; 95%CI : 0.63 to 1.56), Idaho (RR = 0.97; 95%CI : 0.42-1.52), Arkansas (RR = 0.96; 95%CI : 0.64 to 1.29), and Alabama (RR = 0.96; 95%CI : 0.71 to 1.21). Across states, higher vaccination rates were correlated with lower RRs (i.e., faster decline) (rho=-0.42, P = .0027). Every 10% increase in vaccination rates was associated with an 11.5% (95%CI : -17.2% to -5.4%) reduction in RRs, adjusted for screening rates.

Conclusions: Declines in cervical cancer incidence in young women during the HPV vaccination era varied substantially by state, aligning with HPV vaccination rates.

Background: This study examined whether breast cancer survival improvements with screening are explained solely by detection at early stages and whether mortality can be predicted using stage and diagnosis date alone.

Methods: We compared stage-specific net survival between never-screened, symptomatic ever-screened (past attenders and interval cancers), and screen-detected breast cancer cases in Denmark using individual-level electronic health records from January 1, 2010 to December 31, 2022. Lifetables were generated from women without breast cancer, separately for never-screened and ever-screened women. Age-specific all-cause mortality rates were used to calculate excess mortality in women with breast cancer, which was then transformed into net survival.

Results: Of 817 128 women, 32 827 had breast cancer, with 8% presenting as stage III or IV. Survival differences between symptomatic and screen-detected cases were minimal for stages I-III but reached 40% at stage IV, with 5-year net survival for stage IV screen-detected women (74.7%) resembling stage IIIc symptomatic survival in never-screened women (72.6%). Survival from stage IV breast cancer was strongly associated with treatment, with median survival (years) of 4.4-6.0 with surgery, 1.6-2.2 with nonsurgical treatment, and 0.03-0.13 with no treatment; 67% of screen-detected patients received surgery (compared with 23% of never-screened and 27% of symptomatic ever-screened).

Conclusions: Greater survival in screen-detected stage IV cases suggests that breast cancer screening may not have come too late and deserves to be investigated further. Predicting breast cancer mortality using stage at diagnosis and stage-specific survival (without differentiating by route to diagnosis) will underestimate the impact of breast screening on mortality.

Background: Gastric cancer (GC) is often diagnosed at advanced stages, contributing to poor prognosis. Circulating metabolites have emerged as potential biomarkers for GC risk stratification or early detection. We conducted a systematic review of studies investigating the association between metabolites and GC, including both precancerous and cancerous gastric lesions.

Methods: We comprehensively searched PubMed, Embase, and Web of Science for articles published from 2004 to 2025. Eligible studies assessed endogenous metabolites using mass spectrometry- or nuclear magnetic resonance-based platforms in relation to precancerous gastric lesions, GC or GC subtypes. Data were extracted on study design, biospecimen type, analytical approaches, Helicobacter pylori infection, identified metabolites, and model performance.

Results: A total of 52 studies were included, comprising 12 case-only, 31 case-control, five nested case-control, and four cohort studies. Across studies, metabolites reported to differ between GC and non-GC groups and across stages of gastric lesion progression were primarily involved in metabolism of glucose, lipids, amino acids, nucleic acids, and vitamins. Several studies evaluated metabolite-based classification or prediction models, reporting a wide range of performance metrics for distinguishing GC from non-GC conditions and for classifying disease stages. Considerable heterogeneity was observed across studies, limiting direct comparability of findings.

Conclusions: Previous studies have reported associations between metabolites and GC, as well as progression of precancerous lesions, providing insights into gastric carcinogenesis. However, substantial heterogeneity across studies highlights the need for standardized methodological approaches and adjustment for key confounders followed by independent validation and replication in large, well-designed, multi-population studies.

Endometrial cancer (EC) is rising both in incidence and mortality, is involving younger women, and is leading in the US for gynecologic cancer incidence. The application of molecular characterization and targeting treatment to selected molecular types of EC is exemplified by the marked benefit of mismatch repair deficient (dMMR) EC to immune checkpoint inhibitor (ICI) treatment. However, the response to immunotherapy has been less significant in other EC molecular types. We reported previously on the public health relevance of molecular analysis of endometrial cancer types to direct treatment considerations and discussed the limitation in biomarkers predictive of response to immunotherapy or available to examine for treatment selection, outside of mismatch repair deficiency. The current follow-on commentary addresses how new thinking can lead to optimization of immunotherapy applications for endometrial cancer molecular types, how to consider timing and sequencing of immunotherapy with other interventions, and directions for novel immunotherapy combinations. This report outlines key background studies and preclinical observations, directions to overcome inherent resistance, how to leverage ICI to augment clinical response to standard treatments, and considerations for how and when to re-expose patients to ICI treatment(s). The discussions led to potential clinical trial concepts now under development.

Background: Adverse health events have been reported among US service members who served at Karshi-Khanabad (K2) Air Base, Uzbekistan, but prior studies of cancer risk had key limitations.

Methods: We conducted a nested case-control study to assess the association between K2 deployment and cancer outcomes among military personnel deployed between 2001-2005. Eight malignant outcomes (brain, colorectal, liver, urinary tract, pancreatic, prostate, leukemia, non-Hodgkin's lymphoma [NHL]) were identified via health records from the Department of Defense and Veterans Administration through 2022. Cases were matched to controls (1:100) on age, sex, and health system usage. Covariate-adjusted conditional logistic regression models estimated outcome likelihood by exposure surrogates without and with lagging.

Results: Among 619,403 service members, 15,031 (2.3%) were deployed at least once to K2 (median duration 130 days). We identified cancer cases (count; cumulative incidence per 10,000) for brain (450; 7.3), colorectal (1,542; 25.0), liver (113; 1.8), urinary tract (1,831; 29.7), pancreatic (417; 6.8), prostate (5,165; 83.6), leukemia (707; 11.5), NHL (1,700; 27.5). Across seven of eight outcomes, there were no associations between deployment and case status. In contrast, each additional month of deployment increased the odds of NHL by 13% (95% CI, 4%-23%) and deployment >180 days was associated with higher odds (OR 1.78; 95% CI, 1.08-2.94). When lagging exposure by 12-15 years, these associations strengthened (range: 2.38 [1.17-4.82] to 3.09 [1.73-5.52]).

Conclusions: Long-term K2 deployment was associated with increased likelihood of non-Hodgkin's lymphoma. Careful clinical monitoring and continued follow-up of this cohort for cancer and other outcomes are warranted.