Loading evaluation data…
Reliability concern: This paper was retracted on 2024-01-24. See notice: 10.1155/2024/9870248.
Although the implementation of surgery has reduced the mortality of breast cancer, postoperative recurrence is still an important problem bothering patients. DCE-GMRI can not only clearly display the morphological characteristics of breast lesions but also dynamically observe the blood perfusion of the lesions. On account of this, we explored the predictive value of preoperative dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) imaging features in breast cancer patients on postoperative recurrence time of breast cancer. The results showed that DCE-MRI images can clearly show the hemodynamic characteristics and morphological characteristics of tumor lesions, and have important value in predicting the recurrence time of breast cancer after surgery. The prognosis of early recurrence of breast cancer is worse. DCE-MRI can predict the time of postoperative recurrence and provide important clinical references.
Strengths
The study addresses the practical clinical problem of predicting breast cancer recurrence timing using standardized DCE-MRI with both morphological and hemodynamic parameters measured via Omni-Kinetics software, providing a systematic multi-feature assessment.
↳ Introduction; Section 2.4
The Kaplan-Meier analysis with log-rank test (P=0.012) demonstrates that early recurrence is associated with worse survival outcomes (median 26.8 vs. 54.3 months), providing clinically meaningful context for the imaging findings.
↳ Section 3.5
The paper reports concrete quantitative metrics including AUC with 95% confidence intervals and sensitivity/specificity at a specified cutoff, enabling comparison with future studies even if the current performance estimates require external validation.
↳ Section 3.4; Table 3
Limitations
All 86 patients had confirmed postoperative recurrence — no non-recurrent comparator group is included. The ROC analysis (AUC 0.918) is computed within this restricted population, dramatically inflating discriminative performance compared to what would be observed in a general breast cancer population where the clinically relevant question is who will recur.
↳ Section 2.1; Section 3.4
The two groups are compared on DCE-MRI features without controlling for tumor stage, grade, or receptor status (ER/PR/HER2) — all known determinants of both recurrence timing and imaging features. No baseline characteristics table stratified by recurrence group is presented.
↳ Methods Sections 2.4–2.6
Section 2.1 states '66 patients were divided' but the subgroups sum to 86. Section 3.1 reports 'no significant differences' with '(P < 0.05)', directly contradicting the Discussion which correctly states P > 0.05. These errors suggest insufficient quality control in manuscript preparation.
↳ Section 2.1; Section 3.1; Section 4 (Discussion)
Withdrawn — shown for reference only
The adjudicated quality score is driven primarily by fundamental methodological concerns in study design. The restriction of the sample to patients who already recurred is the most consequential limitation: it means the reported AUC of 0.918 cannot be interpreted as clinical prediction performance in a general breast cancer population. The absence of any confounding adjustment for established prognostic factors (tumor stage, grade, receptor status) further limits interpretability. Internal reporting inconsistencies — including a numerical error in the sample description and an inverted P-value statement in the Results — raised concerns about manuscript preparation quality, though these do not invalidate the core directional findings. The paper's context score reflects a limitations section that acknowledges only sample size while omitting the more fundamental design constraints. The impact potential assessment converged around 2.1, reflecting agreement that the work is at an early exploratory stage without a clear path to clinical implementation.
Limited2.3/5.0
The paper addresses a clinically relevant question — whether preoperative DCE-MRI features differentiate early from late breast cancer recurrence — extending prior work from recurrence presence to recurrence timing. However, the advance is incremental given the small single-center sample and the abstract overclaims relative to the exploratory design.
have important value in predicting the recurrence time of breast cancer after surgery
The study is restricted to patients who already recurred (N=86, single center), meaning it cannot assess who will recur — only whether features differ between early and late recurrers. No multivariable adjustment for known confounders (tumor stage, grade, receptor status) is performed, and the ROC model specification is unclear, making the reported AUC of 0.918 uninterpretable as a generalizable predictive metric.
The paper follows a conventional and readable structure with detailed MRI protocol description. However, the inverted P-value statement in Section 3.1 ('no significant differences... P < 0.05') would mislead a reader relying on the Results alone, and the numerical inconsistency in Section 2.1 creates confusion about sample composition.
There were no significant differences in BPE, morphology, lesion margin, intralesional enhancement, and TIC between the two groups... (P < 0.05)
Literature engagement is sparse (12–14 references, many general rather than specific to DCE-MRI recurrence prediction). The limitations section acknowledges only small sample size, omitting the fundamental design limitation of the recurrent-only sample and absence of confounding adjustment.
the shortcoming of this study is that the included samples were small
Concerns4 of 4 checks
The paper contains a numerical inconsistency (Section 2.1 states '66 patients' but subgroup totals sum to 86) and a logical contradiction in statistical reporting (Section 3.1 reports 'no significant differences' with 'P < 0.05', contradicted by the Discussion which correctly states P > 0.05). Additionally, the ROC analysis (Section 3.4) does not specify which variable or combination of variables produced the reported AUC of 0.918.
No ethics approval or informed consent statement is present in the paper, which is a reporting omission for a retrospective clinical study by 2020 standards. No conflict of interest disclosure is provided. The data availability statement is vague, offering data 'without undue reservation' but naming no repository or persistent identifier.
Flags: 1 declared / 5 total
All 14 cited references resolved to real publications via PubMed, Crossref, or OpenAlex.
This paper was retracted on 2024-01-24. See notice: 10.1155/2024/9870248.
Sources: Retraction Watch ✓PubPeer (coming soon)
Low2.1/5.0
The paper targets a clinically relevant problem — predicting breast cancer recurrence timing — with identifiable stakeholders (oncologists, radiologists). However, no specific clinical decision pathway, guideline context, or institutional stakeholder is named.
This is a single-center retrospective observation with an internal-only ROC analysis and no external validation. The work sits at the early exploratory stage, well short of deployment-ready clinical evidence.
Generalizability is severely constrained by the single-center, small-sample design with no sensitivity analyses across scanners, populations, or treatment contexts. Manual ROI delineation is operator-dependent and difficult to standardize.
The paper sits within an active research area (DCE-MRI prognostic imaging in breast cancer) but does not explicitly connect to prior systematic reviews, meta-analyses, or guideline development processes, limiting its contribution to cumulative evidence building.
Lower confidence on Relevance, Trajectory — domain match limited.
Nabu’s assessment, alongside the field’s view.
Evaluation Metadata
AI-powered, human-governed. Something look off? Contact us to request a review.