RCR Category 2 Requires a Positive, Very-High-Confidence Finding of Benignity: Persistence and Developing Microcalcifications Can Mandate Recall

Case: McSweeney v Health Service Executive (Approved) [2026] IEHC 451 (High Court of Ireland, Coffey J, 30 June 2026)
Module tried: Breach of duty only (liability in negligence not finally determined; causation/quantum not addressed in this module).

1. Introduction

The plaintiff, a registered nurse enrolled in a surveillance programme due to a strong family history of breast cancer, alleged negligent interpretation of surveillance mammograms performed in October 2020 and January 2022 at University Hospital Waterford (services overseen by the Health Service Executive). She contended that both studies showed a new cluster of microcalcifications in the upper anterior left breast, that the appearances were indeterminate, and that she should have been recalled for magnification views and biopsy. In 2023, a further surveillance mammogram showed overtly malignant features and biopsy confirmed invasive ductal carcinoma.

The dispute was not about the existence of cancer in 2023, but about whether the subtle features in 2020 and 2022 crossed the professional “recall threshold” under the Royal College of Radiologists (“RCR”) breast imaging classification (particularly the boundary between RCR Category 2 and 3).

2. Summary of the Judgment

• No breach was found in the October 2020 classification as RCR Category 2 (benign). The court held the finding fell within the range of acceptable professional judgment given the low number of visible microcalcifications, their subtlety, plausible benign explanations (including glandular involution/projection effects), and recognised interpretive variability.
• Breach was found in the January 2022 classification as RCR Category 2. By then, the court considered that the cumulative constellation—persistence over time of a focal grouping, increased conspicuity, at least four (possibly five) microcalcifications, and partial alignment on the MLO view—meant that a benign classification could no longer be sustained to the very high level of professional confidence required for Category 2.
• The court concluded that by January 2022 the proper minimum classification was RCR Category 3 (indeterminate/probably benign requiring further investigation), and the failure to recall for assessment at that time constituted breach of duty.

3. Analysis

3.1 Precedents Cited

(a) Dunne v National Maternity Hospital [1989] IR 91

The court treated Dunne as the foundational Irish authority on clinical negligence. It applied the “ordinarily competent practitioner” standard and the principle that, where the issue is one of professional judgment rather than breach of a mandatory rule, negligence is not established merely because another competent practitioner would have acted differently. The key inquiry is whether the defendant’s approach is supported by a responsible body of professional opinion and whether it withstands logical scrutiny.

(b) Morrissey v Health Service Executive [2020] IESC 6

Morrissey was relied upon to clarify and reaffirm Dunne, especially for:

• Hindsight bias control: later adverse outcomes do not prove earlier breach; the court must evaluate what was reasonably open on the evidence available at the time.
• Threshold-based decision-making under variability: where systems require judgments at a threshold (here, RCR 2 vs 3), the existence of interpretive variability matters to what is “reasonably open,” but does not immunise clinicians from scrutiny.
• Logical scrutiny over headcount: a “body of opinion” is relevant but not determinative; the decisive question is whether the conclusion was reasonably open and logically defensible when the methodology is applied to the facts.

Coffey J used Morrissey to reject the plaintiff’s submission that subsequent malignancy meant the earlier appearances “could not have been benign,” characterising that as impermissible hindsight reasoning.

(c) Penney and Ors v East Kent Health Authority [2000] Lloyd's Rep Med 41

The plaintiff sought to draw from this cytology-screening authority the idea that a screener must not classify as normal where there is “any doubt.” The court distinguished it:

• Domain difference: cervical cytology screening involved an operational rule in a binary classification context.
• System difference: this case concerned radiology within a graded RCR system requiring evaluative judgment about confidence thresholds, not a strict “any doubt = abnormal” rule.

The court’s handling of Penney is important: it resists importing domain-specific screening rules into different diagnostic contexts without evidence that the profession applies an equivalent operational rule.

3.2 Legal Reasoning

(a) Framing the legal issue: “Was Category 2 reasonably open?”

The decisive question was not whether recall would have been “safer,” but whether it was reasonably open to competent radiologists to classify as RCR Category 2 at each timepoint. The judgment emphasises that RCR 2 carries a strong clinical implication: the radiologist has a very high degree of confidence that findings are benign such that recall is unnecessary.

(b) The court’s synthesis of the “confidence threshold” for Category 2

Expert evidence used different formulations (“absolute confidence”; “if you have any doubt...”), but the court held these were not legal standards and should not be treated as interchangeable. The judgment synthesised the practical standard as:

This synthesis functioned as the bridge between (i) the RCR system’s clinical meaning and (ii) the Dunne/Morrissey legal test.

(c) Variability is relevant, but not a shield

The judgment accepts both intra-observer and inter-observer variability as “well recognised and documented” in microcalcification interpretation. However, Coffey J stressed that variability does not remove liability: it merely informs the “range” of acceptable professional judgment. The court’s task is to determine whether the defendant’s conclusion remained within that range.

(d) Why 2020 was within the acceptable range

For October 2020 the court found:

• The finding was a very small number of microcalcifications (three, possibly four), and counting at that scale is inherently difficult.
• Morphology was irregular but subtle; it was reasonably open to regard them as amorphous and benign in context.
• No convincing linear/ductal distribution was demonstrated; incidental alignment of very small numbers did not compel recall.
• Interval “new visibility” could plausibly be explained by glandular involution/projection effects in a woman in her forties.

Even if a more cautious approach existed, that did not establish breach under Dunne/Morrissey.

(e) Why 2022 crossed the boundary of permissible judgment

The key move in the reasoning is the cumulative assessment. By January 2022, the court held that the same focal grouping had:

• Persisted over time (no longer plausibly “transient” background change);
• Become more conspicuous;
• Reached at least four (possibly five) microcalcifications within a confined area (potentially meeting a cluster threshold);
• Shown alignment on the MLO projection (not classic linear-branching, but contributing to a qualitative change).

The court treated this as a qualitative shift from possibly incidental dots (2020) to a more coherent focal grouping with persistence and developing number/alignment (2022). Importantly, the court held that the defendant’s justification by 2022 had become essentially: “nothing sufficiently suspicious to mandate recall.” That, the court found, is not enough for Category 2, which requires a positive benign classification to a very high confidence level. On the court’s analysis, the evidence supported “probable benignity” at most—i.e., Category 3.

(f) “Body of opinion” vs “reasonably open”

A notable feature is the court’s willingness to find breach even while accepting that many competent radiologists might have called the 2022 mammogram Category 2. The judgment reiterates that the standard does not turn on numerical weight of opinion; the court must still examine whether the conclusion was logically defensible on the facts when applying a coherent professional methodology. This is a direct application of the “logical scrutiny” component of the Dunne/Morrissey framework.

3.3 Impact

• Clarifies what “benign” means in graded radiology reporting: The decision sharpens the practical boundary between RCR 2 and RCR 3. Where features are explained primarily by “absence of sufficient suspicion” rather than the ability to positively sustain a very-high-confidence benign conclusion, Category 2 may be unsafe in medico-legal terms.
• Encourages explicitly cumulative reasoning: Persistence and development of subtle findings over serial surveillance images can transform the evaluative context, even without dramatic morphological evolution.
• Limits “any doubt” arguments outside their domain: By distinguishing cytology authorities, the judgment discourages transplanting binary-screening rules into graded radiological assessment unless supported by evidence of professional practice.
• Reinforces outcome-neutral analysis: The judgment is a detailed template for avoiding hindsight bias while still allowing courts to find that a threshold was crossed at a later timepoint (2022) even if an earlier call (2020) was defensible.
• Operational implications for services: Although the case was not decided on systems issues, the evidence highlighted differences between “single read” settings and collegial/consensus practice. Future litigation may focus on how the working environment affects what is “reasonably open” when borderline features persist across time.

4. Complex Concepts Simplified

4.1 RCR Categories (in practical terms)

• RCR 2 (Benign): the radiologist is highly confident the finding is harmless; no recall.
• RCR 3 (Indeterminate / probably benign): not overtly suspicious, but not confident enough to call benign; recall for additional imaging and/or assessment.

4.2 Microcalcifications and why “small numbers” are difficult

Microcalcifications are tiny calcium deposits seen as small white dots on mammography. Their significance depends on:

• Number: small numbers can be benign; “clusters” (often described as ≥5) can be more concerning, but counting is difficult when dots are faint.
• Morphology: “amorphous” (hazy) is common and often benign; “pleomorphic” (varied size/shape) is more concerning.
• Distribution: alignment suggesting ductal patterns can be concerning even without classic “linear branching.”
• Interval change/persistence: persistence across serial studies can increase concern even if each single study looks only mildly abnormal.

4.3 Hindsight bias (why the later cancer does not prove earlier negligence)

The court emphasised that early malignant change can look indistinguishable from benign findings on imaging. The legal question is what a competent radiologist could reasonably conclude then, not what is known now.

5. Conclusion

McSweeney v Health Service Executive [2026] IEHC 451 is a structured application of Dunne v National Maternity Hospital and Morrissey v Health Service Executive to graded radiological reporting. It confirms that RCR Category 2 is not justified merely because findings are “not suspicious enough”; it requires a positive, very-high-confidence benign assessment. The case also illustrates how, in surveillance imaging, persistence and developing focal features—even without dramatic progression—may cumulatively require reclassification to RCR Category 3 and recall, and that a court may find breach at a later timepoint even where an earlier non-recall decision was within the permissible range of professional judgment.