Evidence for the prognostic value of TP53 mutations in circulating tumor DNA across solid malignancies: a systematic review and meta-analysis Authors: Lan Jinyan, Liang Yibin, Qin Jiangkui, Huang Shanbo, Glenn Deng, Changsheng Sun, He Yongyu, Huang Gang, Yi Tingzhuang Published: 2026 ABSTRACT BackgroundThe purpose of this meta-analysis study is to provide evidence for the clinical utility of TP53 mutations in circulating tumor DNA (ctDNA) as a prognostic biomarker.MethodsWe searched the PubMed, Embase, Cochrane, and Web of Science databases (last update May 2025) for studies on TP53 mutations in ctDNA or cfDNA as prognosis overall survival and in solid tumors. A total of 21 studies that met the criteria were utilized and data was collected regarding the authors, year of publication, study design, site of the study, number of patients, detection, mutation sample size and outcome measures were collected. The Newcastle-Ottawa Scale (NOS) was used to evaluate the quality of the study, and meta-analysis was done by using STATA 16.0. Effect sizes were in the form of hazard ratios (HR) that had 95% confidence intervals (CI). The models used were fixed-effects and random-effects based on heterogeneity. Funnel plots, and Egger’s test was used to measure publication bias, and sensitivity analysis conducted through a leave-one-out method.ResultsA total of 21 studies (2,685 TP53-mutated patients, one unreported) showed: Mutated patients had worse progression-free survival (PFS) (HR=2.10, p=0.000; 12 studies, heterogeneity resolved after excluding Yoshida 2023), shorter OS (HR=1.74, p=0.014; 9 studies), and reduced DFS (HR=1.73, p=0.007; 3 studies), but RFS (2 items) showed no statistically significant differences. Subgroup analyses revealed: Prospective studies showed stronger PFS (HR=2.14 vs retrospective 1.90) with Japanese subgroup HR=4.90; Lung/liver cancers had higher HRs than breast. Prospective OS HR=2.25 (lung 3.14, endometrial 0.75). Retrospective DFS HR=1.89 vs Japanese breast RFS HR=4.00. Heterogeneity originated from study design, region, and cancer type variations, with no significant publication bias (Egger's test p>0.05).ConclusionCurrent evidence suggests that TP53 mutations detected in ctDNA are significantly associated with poor prognosis in various solid tumors, particularly lung cancer. The association is robust for PFS and OS, though high heterogeneity and biological complexity warrant cautious interpretation. These findings support the potential incorporation of ctDNA-based TP53 mutation status into clinical prognostic assessment systems as an adjunctive parameter; however, further standardization of detection protocols, functional annotation of mutation types (e.g., LOF vs. GOF), incorporation of VAF and clonality analysis, and validation in large prospective multicenter cohorts are needed before routine clinical implementation.PROSPERO registration numberCRD420251021095Supplementary InformationThe online version contains supplementary material available at 10.1186/s12885-026-16265-y. Strengths and limitations of this study In this research, the literature screening was conducted comprehensively, the quality of the studies was evaluated based on the international standards and only those articles that had a low risk of bias were excluded. This meta-analysis confirmed the deleterious prognostic effect of the TP53 mutations using a bigger sample size and stratified analysis and found causes of heterogeneity, such as region, cancer type, and study design, thus proving new evidence on the accurate prognostic measurement. The joint effect of pooled hazard ratio of the progression-free survival was comparable to the findings of the previous single-center studies, but heterogeneity was more enhanced, with large differences in the magnitude of effects among studies. Results on OS were consistent with the current research trends, but there was even more heterogeneity, probably as a result of variations in the study design (prospective and retrospective), or regional differences.Furthermore, this meta-analysis is constrained by the inherent heterogeneity of included studies, including differences in ctDNA collection timing (baseline vs. serial), VAF thresholds, and bioinformatic pipelines. The binary (mutated vs. non-mutated) classification of TP53 does not account for critical biological distinctions—such as missense vs. truncating mutations, loss-of-heterozygosity, or dominant-negative/gain-of-function effects—which may have distinct prognostic implications. The absence of paired white blood cell sequencing in all included studies precludes definitive discrimination of tumor-derived TP53 mutations from clonal hematopoiesis-derived variants. These limitations should be addressed in future prospective studies with harmonized technical standards and biologically informed variant classification. Introduction TP53 gene encodes the p53 protein which acts as a tumor suppressor [1]. TP53 is inactivated by mutations in more than 50% of solid tumors [2]. p53 controls the genomic stability by controlling essential cell mechanisms, such as cell cycle arrest, DNA repair, and apoptosis. Deficiency of p53 results in unchecked tumor cell growth, enhanced metastatic capacity and acquisition of resistance to therapy [3]. Recent studies have elucidated key mechanisms underlying mutp53 stabilization and function. For instance, the cellular redox environment, regulated by glutathione reductase, plays a critical role in maintaining mutp53 stability by antagonizing its glutathionylation and subsequent proteasomal degradation [4]. This redox-mediated stabilization contributes to the heterogeneous accumulation of mutp53 within tumors, which is associated with more aggressive phenotypes and therapeutic resistance. In the era of precision oncology, targeting p53 dysfunction has emerged as a promising therapeutic strategy. Various approaches are being developed to restore p53 function in tumors, including MDM2 and MDM4 small molecules that reactivate mutp53 or inhibit its negative regulators, immunotherapies leveraging p53 neoantigens, and gene therapies delivering wild-type TP53 [5]. These advances underscore the importance of molecular classification and diagnostic stratification of p53-mutated tumors to guide targeted treatment strategies. Over the last few years, the liquid biopsy has emerged as a novel non-invasive approach in keeping track of the mutation status of TP53 by analyzing circulating tumor DNA (ctDNA) and cell-free DNA (cfDNA). such fragments are identifiable in the bodily fluids [6]. ctDNA is a non-cellular fragment of DNA that can be detected in the bodily fluids [7]. Study of ctDNA, which is a subset of cfDNA, on tumor in a liquid biopsy has become a formidable technique of non-invasive cancer surveillance [8]. Nevertheless, limitations associated with the available evidence exist: the sensitivity of different detection platforms differs significantly, and the detection sensitivity of ctDNA varies by mutation type, being higher for KRAS mutations than for others than with other mutations, since the latter ranges between 70 and 100% [9]. It is becoming clear that certain mutant p53 proteins result in more vicious tumor phenotypes, implying the attainment of new oncogenic functions [10, 11].Addressing these challenges is crucial for the clinical utility of ctDNA and enhancing patient care [6]. Hence, to perform a detailed analysis and quantitative synthesis of TP53 mutations in ctDNA as a prognostic biomarker, and their clinical applicability, a systematic review and meta-analysis is warranted. Methods Search strategy The literature search was conducted across the PubMed, Embase, Cochrane, and Web of Science databases, covering publications from the inception of each database to May 2025. The search focused on studies investigating TP53 mutations in ctDNA or cfDNA as prognostic indicators in various solid tumors. Both subject headings and free-text terms were used, including: “Cell-Free Nucleic Acids,” “cell-free DNA,” “cfDNA,” “circulating tumor DNA,” “ctDNA,” “liquid biopsy,” “plasma DNA,” “serum DNA,” “cfDNA analysis,” “ctDNA analysis,” “Genes, p53,” “TP53 mutation,” “TP53,” “p53 gene,” “mutations in TP53,” “p53 gene mutations,” “TP53 mutation status,” “p53 alterations,” “p53 mutations,” “solid tumors,” “non-small cell lung cancer,” “NSCLC,” “breast cancer,” “lung cancer,” “colorectal cancer,” “gastric cancer,” “hepatocellular carcinoma,” “prostate cancer,” “ovarian cancer,” “esophageal cancer,” “pancreatic cancer,” and “endometrial cancer.” The study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [12], which provide evidence-based standards to enhance the methodological quality of systematic reviews containing primary data. The study protocol was registered with PROSPERO (Registration No. CRD420251021095) [13]. The PRISMA checklist and detailed PROSPERO registration information are provided in Supplemental Files 1 and 2, respectively. The primary endpoint of this systematic review was to assess the association between TP53 mutation status in ctDNA and prognosis, including OS and PFS, in patients with solid tumors. The initial search retrieved 3,449 records, and their abstracts were screened for relevance. The inclusion criteria for this study were as follows: Inclusion and exclusion criteria Include studies that meet the following criteria: The study population comprised patients with histologically confirmed solid tumors harboring TP53 mutations detected in ctDNA;Studies reported at least one of the pre-specified survival endpoints (PFS, OS, DFS, or RFS);No requirement for tissue confirmation of TP53 mutations, but concordance data were recorded when available. The study population comprised patients with histologically confirmed solid tumors harboring TP53 mutations detected in ctDNA; The study population comprised patients with histologically confirmed solid tumors harboring TP53 mutations detected in ctDNA; Studies reported at least one of the pre-specified survival endpoints (PFS, OS, DFS, or RFS); Studies reported at least one of the pre-specified survival endpoints (PFS, OS, DFS, or RFS); No requirement for tissue confirmation of TP53 mutations, but concordance data were recorded when available. No requirement for tissue confirmation of TP53 mutations, but concordance data were recorded when available. Studies will be excluded if any of the following conditions are met: Patients with solid tumors but TP53 mutations not detected in ctDNA;Studies who’s primary or reported outcomes did not include at least one of the pre-specified survival endpoints (i.e., PFS, OS, DFS, or RFS);Reviews, case studies, survey analyses, dissertations, conference abstracts, and irrelevant literature;Duplicate publications. Patients with solid tumors but TP53 mutations not detected in ctDNA; Patients with solid tumors but TP53 mutations not detected in ctDNA; Studies who’s primary or reported outcomes did not include at least one of the pre-specified survival endpoints (i.e., PFS, OS, DFS, or RFS); Studies who’s primary or reported outcomes did not include at least one of the pre-specified survival endpoints (i.e., PFS, OS, DFS, or RFS); Reviews, case studies, survey analyses, dissertations, conference abstracts, and irrelevant literature; Reviews, case studies, survey analyses, dissertations, conference abstracts, and irrelevant literature; Duplicate publications. Duplicate publications. Literature screening and data extraction Two reviewers independently screened the titles, abstracts, and full texts of all retrieved studies to determine eligibility based on the predefined inclusion criteria. In cases of disagreement, the original articles were re-examined, and consensus was reached through discussion. Data were extracted from the included studies, including authors, publication year, study design, study location, patient population, ctDNA analysis methods, TP53 mutation sample size, and reported outcomes. These data were then compiled into a structured three-column table. The quality of included case-control studies was assessed independently by the two reviewers using the NOS, which evaluates nine dimensions: selection of the exposed group, selection of the non-exposed group, ascertainment of exposure, confirmation that the outcome of interest was not present at baseline, control for confounding factors, comparability of study and control groups, accuracy and objectivity of outcome assessment, adequacy of follow-up duration, and completeness of follow-up. Studies scoring 1–3, 4–6, and 7–9 points were classified as low, medium, and high quality, respectively [14]. Statistical analysis Data extracted from the included studies were analyzed using Stata 16.0 for meta-analysis. The analysis process was as follows: (1) Effect size: HRs with 95% CIs were used to estimate effect sizes. (2) Heterogeneity assessment: Heterogeneity among studies was evaluated using the P-value and I² statistic. Low heterogeneity was defined as P > 0.1 and I² ≤ 50%, in which case a fixed-effects model (FE) was applied. High heterogeneity was defined as P ≤ 0.1 and I² > 50%, for which a random-effects model (RE) was used. (3) Publication bias: Funnel plots and Egger’s test were employed to assess potential publication bias for PFS and OS outcomes. (4) Sensitivity analysis: Stepwise exclusion (leave-one-out) methods were applied to PFS and OS to evaluate the robustness of the results. Technical considerations of TP53 mutation detection All included studies utilized next-generation sequencing (NGS) for ctDNA analysis, but the genomic coverage of TP53 varied across studies. As shown in Supplemental File 3, eTable 5, some studies targeted only hotspot regions (exons 5–8), while others covered the entire coding sequence. This technical heterogeneity, along with differences in sequencing depth, bioinformatics pipelines, and variant calling criteria, may affect mutation detection sensitivity and comparability. We have summarized these technical details insupplemental file 4 (newly added to Supplemental File 4) to facilitate transparency. Results Literature search results A total of 4,285 records were initially identified, of which 3,449 remained after removing duplicates. Following a review of titles, abstracts, and full texts, and based on strict adherence to the predefined inclusion and exclusion criteria, 21 studies were ultimately included [15–33]. The detailed search strategy for each database is provided in Supplemental File 3, along with eTables 1–4. The study selection process is illustrated in Fig. 1. Fig. 1Flow chart of literature screening according to PRISMA guidelines. The process of selecting articles for title and abstract based on inclusion exclusion criteria.The initial search retrieved 4,285 articles, with 3,449 retained after removing duplicates. Through reviewing titles, abstracts, and full texts while strictly adhering to inclusion and exclusion criteria, 21 studies were ultimately included Flow chart of literature screening according to PRISMA guidelines. The process of selecting articles for title and abstract based on inclusion exclusion criteria.The initial search retrieved 4,285 articles, with 3,449 retained after removing duplicates. Through reviewing titles, abstracts, and full texts while strictly adhering to inclusion and exclusion criteria, 21 studies were ultimately included Included study characteristics results A total of 20 studies included 2,685 TP53-mutated patients, with one study (Enric Domingo 2018) not reporting the exact number of patients. Detailed patient numbers for each study are provided in Supplemental File 5.For the Domingo 2018 study, the HR for PFS was based on the presence of any TP53 mutation in ctDNA, though the precise mutation-positive sample size could not be extracted; this study was therefore included in the meta-analysis but excluded from the mutation count summary.The studies were published between 2018 and 2024. Among them, eight were retrospective and twelve were prospective studies, with Enric Domingo 2018 incorporating both prospective and retrospective designs. Study locations included Lithuania, China, Japan, the United States, South Korea, and global multicenter sites. The patient populations encompassed lung cancer, liver cancer, colorectal cancer, prostate cancer, breast cancer, and endometrial cancer. All ctDNA analyses were performed using next-generation sequencing (NGS) methods. Reported outcomes included PFS, OS, disease-free survival (DFS), and recurrence-free survival (RFS). Detailed characteristics of the included studies are provided in Supplemental File 3, eTable 5. Literature quality evaluation results The quality of individual studies was independently evaluated by two reviewers (HY and HS) using a modified NOS [34]. Studies were classified as having high (< 5 stars), moderate (5–7 stars), or low (≥ 8 stars) risk of bias. Among the included studies, six were rated 5 points, four were rated 6 points, one was rated 7 points, and ten were rated 8 points, indicating overall acceptable study quality (see Supplemental File 3, eTable 6). Meta-analysis results The impact of TP53 mutations in ctDNA on patients’ PFS.Twelve studies reported PFS The heterogeneity test indicated significant variability among studies (p < 0.001, I² = 79.3%) (Supplemental File 3, eFigure 1-k-l), and therefore a random-effects model was applied for pooling effect sizes. Meta-analysis results showed that patients with TP53 mutations in ctDNA had shorter PFS compared to those without TP53 mutations [HR = 2.10, 95% CI: 1.58–2.80, Z = 5.087, p < 0.001] (Supplemental File 3, eFigure 1-k-l). Visual inspection of the funnel plot revealed a slight asymmetry, with more data points on the left than the right. Points on the left clustered mainly in the upper-middle region, while those on the right were primarily in the lower-middle region. Three study data points fell outside the 95% confidence interval. However, Egger’s test (p = 0.744) indicated a low likelihood of publication bias (Supplemental File 3, eFigure1-g-h). Sensitivity analysis was conducted using a leave-one-out approach, sequentially excluding each study and re-analyzing the remaining studies. A sensitivity analysis was conducted by excluding the study by Tatsuya Yoshida 2023, which had the highest effect estimate (HR = 4.90). When the remaining 11 studies were pooled using a fixed-effect model, the overall HR was 1.70 (95% CI: 1.49–1.93; z = 8.033, p < 0.001;Supplemental File 3, eFigure 1-n). Heterogeneity was no longer present (I² = 0.0%, p = 0.624,Supplemental File 3, eFigure 1-m), indicating that the high heterogeneity in the main analysis was largely driven by this single outlying study. The direction and significance of the association remained consistent.Accordingly, the results of this study should be interpreted with caution, and the overall PFS association remains statistically significant even after its exclusion (Supplemental File 3, eFigure 1-m). The impact of TP53 mutations in ctDNA on OS There are nine independent studies that presented information about the general survival rate. The meta-analysis showed that the heterogeneity of the studies was significant (p < 0.001, I² = 87.9%) (Supplemental File 3, eFigure 2-k-l). The extent of inconsistency in each of the studies necessitated the application of a random-effects model to the issue of pooling the effect sizes. In doing the meta-analysis with this model it was found that patients who had TP53 mutations in ctDNA had an OS that was significantly lower than that of patients without these mutations. The calculated hazard ratio (HR) to use in this comparison was 1.74, with the 95% confidence interval (CI) of 1.12 to 2.71. The statistical significance of this result was further confirmed by a Z-value of 2.460 and a corresponding P-value of 0.014, suggesting that the observed difference in survival outcomes between the two groups is unlikely to have occurred by chance alone (Supplemental File 3, eFigure 2-l). The funnel plot showed an asymmetric distribution of study data points, with more points on the right than on the left, and four study data points lying outside the 95% confidence interval. However, Egger’s test (p = 0.472) suggested a low probability of publication bias (Supplemental File 3, eFigure 2-g-i). Sensitivity analysis revealed that all point estimates of the included studies for the OS indicator fell within the 95% CI of the pooled effect size, indicating good stability of the study results (Supplemental File 3, eFigure 2-j). The impact of TP53 mutations in ctDNA on Disease-Free Survival(DFS) and Recurrence-Free Survival (RFS) Three studies reported DFS rates. The heterogeneity test (p = 0.045, I² = 67.8%) indicated significant heterogeneity; thus, a random-effects model was employed for effect size pooling. Meta-analysis results showed that patients with TP53 mutations in ctDNA had shorter DFS compared to those without TP53 mutations [HR = 1.73, 95% CI (1.16, 2.58), Z = 2.699, p = 0.007] (Supplemental File 3, eFigure3-a-d). Two studies reported RFS. The heterogeneity test (p = 0.078, I² = 67.8%) indicated significant heterogeneity; therefore, a random-effects model was used to pool the effect sizes. Meta-analysis results showed no statistically significant difference in RFS between patients with TP53 mutations in ctDNA and those without TP53 mutations [HR = 1.93, 95% CI (0.67, 5.53), Z = 1.226, p = 0.220] (Supplemental File 3, eFigure 4-a-b). Subgroup analysis of the predictive role of TP53 mutations in ctDNA for prognosis across various solid tumors In subgroup analyses (Supplemental File 3, eTable 7), the prognostic predictive effect of TP53 mutations in ctDNA for various solid tumors exhibited significant heterogeneity. The overall PFS analysis showed that TP53 mutations were significantly associated with poor prognosis (HR = 2.10, 95% CI 1.58–2.80, p < 0.001), but with high inter-study heterogeneity (I² = 79.3%). By tumor type, TP53 mutations were more strongly associated with prognosis in lung cancer (HR = 2.29, p = 0.007) and liver cancer (HR = 2.73, p = 0.001) patients, whereas the association was relatively lower in breast cancer (HR = 1.66, p < 0.001) (Supplemental File 3, eFigure 1-a-b). Among studies from different regions, the Japanese subgroup had the highest hazard ratio (HR = 4.90, 95% CI 3.70–6.49, p < 0.001), while the association was relatively weaker in the Chinese subgroup (HR = 1.55, 95% CI 1.28–1.87, p < 0.001), and the Lithuanian subgroup (HR = 2.50, 95% CI 0.81–7.76, p = 0.113) showed no statistical significance (Supplemental File 3, eFigure 10). When stratified by study type, the hazard ratio was higher in prospective studies (HR = 2.14, 95% CI 1.52–3.02, p < 0.001) than in retrospective studies (HR = 1.90, 95% CI 1.15–3.15, p = 0.012) (Supplemental File 3, eFigure 1-c-d). The overall OS analysis demonstrated that TP53 mutations were significantly associated with poor prognosis (HR = 1.74, 95% CI 1.12–2.71, p = 0.014), but with high inter-study heterogeneity (I² = 87.9%). Among different cancer types, TP53 mutations showed no association with poor prognosis in breast cancer (HR = 1.73, p = 0.131), while the prognostic risk decreased sequentially in lung cancer (HR = 3.14, p < 0.001), prostate cancer (HR = 1.58, p < 0.001), and endometrial cancer (HR = 0.75, p = 0.021) (Supplemental File 3, eFigure 2-a-b.). When stratified by study location, studies from China (HR = 1.57, 95% CI 0.63–3.88, p = 0.333) and global multicenter studies (HR = 1.46, 95% CI 0.76–2.79, p = 0.252) suggested no significant link between TP53 mutations and poor prognosis, whereas studies from Lithuania (HR = 3.40, p = 0.022) and South Korea (HR = 2.61, p = 0.015) indicated significant correlations (Supplemental File 3, eFigure 2-c-d). Subgroup analysis by study type revealed no correlation between TP53 mutations and poor prognosis in retrospective studies (HR = 1.12, 95% CI 0.40–3.15, p = 0.830), whereas prospective studies showed a significant association (HR = 2.25, 95% CI 1.61–3.15, p < 0.001) (Supplemental File 3, eFigure 2-e-f). The overall DFS analysis showed that TP53 mutations were significantly associated with poor prognosis (HR = 1.73, 95% CI 1.16–2.58, p = 0.007)(Supplemental File 3, eFigure 3-c-d), but with high inter-study heterogeneity (I² = 68.0%). All included studies were conducted in China and focused on breast cancer. Subgroup analysis by study type revealed no correlation between TP53 mutations and poor prognosis in prospective studies (HR = 1.85, 95% CI 0.81–4.22, p = 0.146), whereas retrospective studies demonstrated a significant association (HR = 1.89, 95% CI 1.31–2.72, p = 0.001) (Supplemental File 3, eFigure 3-a-b). In the RFS subgroup analysis, the overall results did not show statistical significance (HR = 1.93, 95% CI 0.67–5.53, p = 0.220), with high heterogeneity (I² = 67.8%), indicating substantial variation across studies (Supplemental File 3, eFigure 4-a-b). However, subgroup analyses revealed the following specific associations: TP53 mutations were significantly linked to increased RFS risk in retrospective studies/Japan/breast cancer (HR = 4.00, 95% CI 1.20–13.90, p = 0.028), while a lower but still significant risk was observed in prospective + retrospective/global multicenter/prostate cancer studies (HR = 1.30, 95% CI 1.01–1.67, p = 0.039). The lack of significance in the overall effect may be attributed to the dilution of subgroup-specific associations by high heterogeneity. Discussion Previous studies have generally found that TP53 mutations in ctDNA are associated with a poor prognosis in patients with solid tumors [23]. This meta-analysis shows that the combined HR for PFS is 2.10 (95% CI 1.58–2.80), consistent with previous single-center studies (e.g., HR = 1.8–2.5), but with high heterogeneity (I² = 79.3%), indicating significant differences in effect sizes across studies. Sensitivity analysis identified the study by Tatsuya Yoshida 2023 as the primary source of heterogeneity for PFS. After its exclusion, heterogeneity was eliminated (I² = 0%), and the pooled HR remained significant at 1.70 (95% CI: 1.49–1.93), indicating a robust association albeit with a slightly attenuated effect size.The combined HR for OS is 1.74 (95% CI 1.12–2.71), consistent with existing studies (HR ≈ 1.5–2.0), but with even higher heterogeneity (I² = 87.9%), possibly due to differences in study design (prospective vs. retrospective) or regional variations. The combined HR for DFS is 1.73 (95% CI 1.16–2.58), consistent with breast cancer-related studies (HR = 1.5–2.0); however, given that only Chinese breast cancer patients were included, caution is advised when generalizing the findings. Furthermore, our meta-analysis synthesizes evidence based on the detection of TP53 mutations in ctDNA [35], which are inherently somatic alterations acquired by tumor cells. This distinguishes them from germline TP53 mutations, which are heritable and present in all cells of an individual, conferring a greatly increased lifetime risk of developing multiple cancers, as in Li-Fraumeni syndrome. The somatic TP53 mutations detected in the included studies are pivotal drivers of clonal evolution. The inactivation of TP53, a key guardian of the genome, removes a critical barrier to genomic instability, allowing tumor cells to accumulate further genetic alterations, thereby fueling tumor progression, heterogeneity, and adaptation to therapeutic pressures. This evolutionary process is exemplified in resistance mechanisms, such as the transformation of EGFR-mutant lung adenocarcinoma to small-cell lung cancer [36], which is nearly universally underpinned by the co-inactivation of TP53 and RB1, enabling the emergence of an aggressive, therapy-resistant clone. Our study suggests that the prognostic significance of TP53 mutations in ctDNA should be interpreted in the context of dynamic clonal evolution. Their presence in circulation likely signals the dominance of a subclone that has leveraged the loss of p53 function for survival advantage, disease progression, and, potentially, the development of treatment resistance. It should be noted that clonal hematopoiesis (CH) may represent a potential confounding factor in the interpretation of TP53 mutations detected in cfDNA. Somatic mutations in genes such as TP53 can arise in hematopoietic stem cells as part of age-related clonal expansion and may be released into the circulation independently of the tumor [37]. None of the included studies performed matched sequencing of white blood cells to definitively distinguish tumor-derived mutations from those originating from CH. Therefore, a proportion of TP53 mutations identified in “ctDNA” could in fact reflect CH-related variants rather than true tumor-derived alterations, which may lead to overestimation of mutation prevalence or prognostic associations in some cases [38]. Moreover, the variant allele frequency (VAF) of TP53 mutations—a critical parameter for distinguishing clonal dominance from subclonal heterogeneity—was not consistently reported across studies. In tumors where TP53 mutation is a truncal (clonal) event, VAF is typically high (approaching the maximum somatic VAF in plasma), whereas subclonal or late-acquired TP53 mutations have lower VAF and may carry different prognostic implications. The failure to integrate VAF and clonality information, along with the absence of CHIP filtering, represents a major source of noise that could either dilute or inflate the true prognostic signal. Future prospective studies should incorporate paired white blood cell sequencing (or buccal swab controls) and report VAF thresholds for defining “positive” ctDNA status, ideally distinguishing clonal from subclonal alterations. This is particularly important for TP53, given its frequent involvement in age-related CHIP. A major limitation of the current evidence base—and consequently of this meta-analysis—is the reliance on a binary “mutated vs. non-mutated” classification of TP53, which obscures substantial biological heterogeneity. TP53 mutations are not functionally equivalent. In clinical oncogenetics, interpretation goes far beyond simple presence or absence; it requires detailed assessment of variant type, allelic status, and functional consequences. First, missense mutations (particularly in the DNA-binding domain) often produce a stable mutant protein that can exert dominant-negative effects (DNE) over the wild-type allele or acquire neomorphic gain-of-function (GOF) properties, promoting metastasis and chemoresistance. In contrast, truncating mutations (nonsense, frameshift, splice-site) typically result in complete loss-of-function (LOF) through nonsense-mediated mRNA decay or protein truncation, often requiring biallelic inactivation (via loss of heterozygosity, LOH) to fully eliminate p53 activity. Second, allelic status is critical: TP53 is a classic tumor suppressor requiring biallelic inactivation. A missense mutation with retention of the wild-type allele may have different functional consequences compared to the same mutation with LOH. Third, the distinction between LOF, DNE, and GOF has distinct prognostic implications—GOF mutations (e.g., R175H, R248W, R273H) have been associated with more aggressive phenotypes and worse outcomes compared to LOF mutations in some solid tumors. None of the included studies stratified TP53-mutated patients by these functional categories, and therefore our pooled HRs represent an average effect across biologically heterogeneous subgroups. This limitation is analogous to the historical pitfalls of p53 immunohistochemistry (IHC), where antibody-based detection of stabilized mutant p53 protein did not reliably reflect underlying mutational status or functional impact. Until primary studies report variant-level outcomes, meta-analyses like ours cannot resolve this granularity. We strongly encourage future ctDNA studies to adopt a functionally annotated classification of TP53 mutations (e.g., LOF vs. DNE/GOF, with allelic context) to enable more precise prognostic stratification. Methodologically, to maximize literature retrieval, our search strategy included both “ctDNA” and “cfDNA.” However, the conclusions of this meta-analysis specifically refer to ctDNA, that is, the portion of cell-free DNA derived from tumors and carrying somatic mutations such as TP53 [39]. This distinction is crucial, as it underscores the direct biological significance of tumor-specific variants detected by liquid biopsy for prognostic evaluation [4]. Additionally, based on the NOS assessment, the overall quality of the included studies was acceptable (most scored between 5 and 8 stars), indicating moderate to high methodological rigor. However, differences in study design (prospective vs. retrospective), follow-up duration, and adjustments for confounding factors may have contributed to the observed heterogeneity. The higher hazard ratios observed in prospective studies compared to retrospective ones may not only reflect design differences but also the typically more rigorous and standardized follow-up protocols in prospective studies. To investigate the sources of heterogeneity, we performed sensitivity and subgroup analyses. These analyses identified study design as a key factor: the hazard ratio for PFS was higher in prospective studies (HR = 2.14) than in retrospective studies (HR = 1.90), suggesting that retrospective studies may underestimate the prognostic impact of TP53 mutations. Our subgroup analysis revealed a notably higher hazard ratio (HR = 4.90) associated with TP53 mutations in the Japanese population compared to other regions. Specifically, lung cancer (HR = 2.29) and liver cancer (HR = 2.73) showed a stronger PFS correlation, whereas breast cancer (HR = 1.66) showed a weaker association. This intriguing finding may be attributed to a confluence of factors. In terms of genetic background, specific genetic modifiers or a distinct prevalence of certain TP53 mutation types within the Japanese population could amplify the negative prognostic impact of these mutations. A comprehensive transcriptome analysis across multiple cancer types suggested that while TP53 mutation-associated biology might be broadly consistent, the influence of specific mutation types could vary. Heterogeneity in treatment effects is a critical consideration in clinical research. The substantially higher HR may reflect differential responses to specific therapeutic protocols more commonly administered in the Japanese healthcare context. The effectiveness of a given treatment can be modulated by individual patient characteristics, and the same principle may apply to population-level responses when stratified by genetic markers. If first-line treatments in the included Japanese studies differed from those in other regions, and if the efficacy of these treatments is particularly compromised by TP53 mutations, this could manifest as an elevated HR. From a technical and methodological perspective, while all included studies utilized next-generation sequencing methodologies, subtle differences in testing platforms, ctDNA extraction protocols, bioinformatic pipelines, or variant interpretation criteria across studies from different regions could introduce variability. Temporal heterogeneity in ctDNA assessment represents another critical source of variability across included studies. The timing of liquid biopsy collection—whether at baseline (pre-treatment), during treatment, or at progression—varies substantially among studies. For instance, some studies assessed TP53 mutations only at diagnosis or before therapy initiation [e.g., 19,27], while others performed serial monitoring at multiple timepoints [e.g., 15,26,32]. In studies with longitudinal sampling, a patient could be mutation-positive at one timepoint and negative at another, yet the overall mutation status used for prognostic analysis was often defined inconsistently (e.g., baseline positivity only, any-time positivity, or persistent positivity). Such methodological heterogeneity directly impacts the interpretation of HRs: a single baseline assessment may miss emerging mutations associated with acquired resistance, while any-time positivity may dilute the prognostic signal due to transient or subclonal mutations. Furthermore, the distinction between pre-treatment, on-treatment, and post-progression ctDNA analysis has profound biological implications. Pre-treatment TP53 mutations reflect the intrinsic clonal architecture of the tumor, whereas on-treatment emergence of new TP53 mutations or an increase in variant allele frequency (VAF) may signal clonal evolution and therapeutic resistance. Unfortunately, the majority of included studies did not stratify outcomes by these temporal variables, precluding a more nuanced meta-analysis. We recommend that future studies standardize reporting of ctDNA collection timing and, where feasible, assess dynamic changes in TP53 VAF rather than relying on a static binary classification. This would significantly reduce heterogeneity and enhance the clinical utility of ctDNA-based TP53 monitoring. The differences between this study and previous Meta analyses. Sample coverage: This study includes 12 PFS studies and 9 OS studies, which is a larger sample size compared to earlier Meta analyses that only included 5 OS studies [40]. However, there are only 2 RFS studies, and the conclusions need further validation. In terms of bias control: This study uses the Egger test (p = 0.744) and sensitivity analysis (for example, by excluding Tatsuya Yoshida’s 2023 study, which reduced the I² to 0%) to identify sources of heterogeneity and reduce bias risk, outperforming previous studies that did not adequately analyze heterogeneity [41]. It is worth mentioning that this study focuses on solid malignant tumors, as there are fundamental differences between solid tumors and hematologic malignancies in terms of biological origin, microenvironment, and mechanisms of circulating nucleic acid release [42]. ctDNA from solid tumors primarily originates from apoptosis or necrosis at the lesion site, and its detection and interpretation systems have established relatively independent clinical pathways [43]. In contrast, molecular monitoring of hematologic malignancies often emphasizes different analytes and genomic markers, with liquid biopsy strategies significantly differing from those for solid tumors. Combining the two for analysis would introduce high heterogeneity, which is detrimental to accurately assessing the prognostic value of TP53 mutations in specific tumor types. Therefore, to maintain research consistency and clinical relevance of the conclusions, this meta-analysis is limited to the scope of solid tumors. Future similar studies targeting hematologic malignancies will contribute to a more comprehensive understanding of the significance of TP53 mutations in different cancer contexts. Moreover, this study has several points of controversy and limitations. Firstly, RFS (relapse-free survival) was not statistically significant: the combined HR (hazard ratio) was 1.93 (p = 0.220), which contradicts some single-center studies, such as the Japanese breast cancer study with an HR of 4.00. This discrepancy may be due to insufficient sample size or subgroup heterogeneity masking the true effect. Secondly, there are contradictions in subgroup analysis: in the OS (overall survival) analysis, the Chinese subgroup (HR = 1.57) and the global multicenter subgroup (HR = 1.46) showed no significant association, whereas the Korean (HR = 2.61) and Lithuanian (HR = 3.40) subgroups showed significant associations, suggesting that regional differences or treatment regimens need further standardization. Importantly, the clinical applicability of our findings should be interpreted with caution. The current evidence is predominantly based on retrospective studies and derived using heterogeneous detection platforms, which may introduce bias and limit the generalizability of the pooled estimates. Although this meta-analysis confirmed the negative impact of TP53 mutations on the prognosis of various solid tumors through a larger sample size and stratification methods, it also revealed various forms of heterogeneity.Our findings support the potential utility of ctDNA-based TP53 mutation monitoring in prognostic assessment. However, several steps are needed before clinical adoption: (1) standardization of pre-analytical and analytical procedures; (2) validation in prospective, multicenter cohorts; (3) integration with other biomarkers and clinical parameters. While our study focuses on prognostic value, liquid biopsy may also have potential applications in early recurrence detection and treatment response monitoring, though these require further investigation in dedicated studies. Conclusions Current evidence suggests that TP53 mutations detected in ctDNA are significantly associated with poor prognosis in various solid tumors, particularly lung cancer. The association is robust for PFS and OS, though high heterogeneity and biological complexity warrant cautious interpretation. These findings support the potential incorporation of ctDNA-based TP53 mutation status into clinical prognostic assessment systems as an adjunctive parameter; however, further standardization of detection protocols, functional annotation of mutation types (e.g., LOF vs. GOF), incorporation of VAF and clonality analysis, and validation in large prospective multicenter cohorts are needed before routine clinical implementation. Supplementary Information Supplementary Material 1. Supplementary Material 1. Supplementary Material 2. Supplementary Material 2. Supplementary Material 3. Supplementary Material 3. Supplementary Material 4. Supplementary Material 4. Supplementary Material 5. Supplementary Material 5.