# Magnetic Resonance Imaging Features in Diagnosis of Breast Cancer and Evaluation of Effect of Epidermal Growth Factor Receptor-Targeted Therapy.

> **Reliability concern:** This paper was retracted on 2024-01-09. See notice: 10.1155/2024/9814185.

- **DOI:** [10.1155/2022/3127058](https://doi.org/10.1155/2022/3127058)
- **Rating:** Limited (2.1 out of 5)
- **Verdict:** A single-centre breast cancer imaging study whose segmentation conclusions contradict its own metrics and whose treatment claims lack the randomisation they assert.
- **Framework version:** v5.0
- **Pipeline version:** v0.4
- **Evaluation date:** 2026-06-01
- **Evaluation page:** https://www.nabu.science/paper/10.1155%2F2022%2F3127058

## Strengths

- **Moderately large single-centre clinical cohort** The study includes 858 patients with pathological confirmation, providing a reasonable sample for a single-centre investigation of both MRI diagnostic parameters and treatment response monitoring.
  - Dimension: contribution
  - Location: Section 2.1
- **Detailed MRI parameter reporting across multiple measures** The paper reports multiple quantitative MRI outcomes (TIC parameters, semiquantitative measures, ADC values, tumour diameter changes) with explicit numerical results for both diagnostic differentiation and treatment response, following established breast MRI methodology.
  - Dimension: methodological rigour
  - Location: Sections 2.3, 3.2–3.5
- **ROC analysis with diagnostic cut-offs for treatment monitoring** The paper provides ROC curves with area-under-curve values, cut-off points, sensitivity, and specificity for multiple MRI-derived parameters (ΔD%, ΔADC%, ΔSImax%, ΔpH%) as treatment response indicators.
  - Dimension: methodological rigour
  - Location: Sections 3.3–3.5, Figures 8, 10, 13

## Limitations

- **Segmentation conclusion contradicts reported metrics** The paper concludes machine segmentation has 'higher accuracy' than manual segmentation, but 4 of 5 reported evaluation metrics (RVD: 6.82% vs 4.63%; VOE: 10.27% vs 6.36%; ASSD: 1.47mm vs 0.98mm; MSSD: 3.88mm vs 1.85mm) show manual segmentation performing better. Only RMS favours machine segmentation.
  - Dimension: reporting
  - Location: Section 3.1, Discussion paragraph 1
- **Treatment comparison lacks randomisation and confounding control** Despite claiming patients were 'randomly divided,' the methods describe allocation based on 'subsequent treatment modalities.' No randomisation procedure, allocation concealment, baseline characteristics comparison (beyond tumour size), or confounding adjustment strategy is described, yet the paper draws causal conclusions about EGFR-targeted therapy superiority.
  - Dimension: methodological rigour
  - Location: Sections 2.1, 3.3–3.5, Discussion, Conclusions
- **Unspecified treatment protocol limits interpretability** The EGFR-targeted agent is never named, the dose, duration, and line of therapy are not described, and patient selection criteria for EGFR therapy (e.g., EGFR expression status, breast cancer subtype) are not reported. This makes the therapeutic findings uninterpretable and non-replicable.
  - Dimension: methodological rigour
  - Location: Section 2.1, Discussion

## Reliability checks

Concerns (4 of 4 checks).

- **Internal coherence:** Concern. Multiple internal contradictions are present: the segmentation accuracy conclusion contradicts 4 of 5 reported metrics, the paper claims 'random division' but describes treatment allocation by clinical modality, parameter counts in benign/malignant data are inconsistent with parameter labels, and the SSmax directional interpretation in the Discussion contradicts the Results section data.
- **Research conduct:** Noted. Ethics approval is declared but no approval number is provided. Informed consent is stated. No conflict of interest statement, no trial registration, and no power calculation are present. Data availability is claimed upon reasonable request with no repository or persistent identifier.
- **Reference integrity:** Clean. All 30 cited references resolved to real publications via PubMed, Crossref, or OpenAlex.
  - References not verified automatically are resolved manually within 24 hours.

## Post-publication record

- **Status:** Concern
- **Summary:** This paper was retracted on 2024-01-09. See notice: 10.1155/2024/9814185.
- **Last checked:** 2026-10-17
