High Court finds Health Service Executive negligent for misclassifying mammograms as benign when they were indeterminate.

THE HIGH COURT

[2026] IEHC 451

[Record No. 2024/5690 P]

BETWEEN

ÁINE MCSWEENEY

PLAINTIFF

-AND-

HEALTH SERVICE EXECUTIVE

DEFENDANT

Judgment of Mr Justice Paul Coffey delivered on the 30th day of June 2026

Introduction

1. The plaintiff is a registered nurse who, at all material times, participated in periodic breast imaging owing to a strong family history of breast cancer.

2. The defendant is the statutory body responsible for the provision and oversight of public hospital services in the State, including diagnostic breast imaging services at University Hospital Waterford.

3. In these proceedings, the plaintiff alleges negligence by the defendant in the interpretation of mammograms performed in October 2020 and January 2022, resulting, she contends, in a failure to recall her for further assessment, delayed biopsy, and delayed diagnosis of breast cancer, which was confirmed in 2023.

4. The plaintiff contends that each of the 2020 and 2022 mammograms demonstrated a new cluster of microcalcifications in the upper anterior left breast, that the morphology and distribution of those microcalcifications were indeterminate, that their indeterminacy mandated recall and that the failure to recall her for magnification views and biopsy amounted to a breach of duty.

5. By agreement between the parties, the trial proceeded on a modular basis, with the present module confined to the issue of breach of duty.

Background

The RCR Classification System

6. As this issue turns on the application of the Royal College of Radiologists (‘RCR’) classification system, it is necessary to explain that system at the outset.

7. Mammograms in Ireland are commonly reported using the RCR breast imaging classification, a five-point scale intended to communicate the radiologist’s level of concern.

8. The categories are as follows:

• RCR Category 1: Normal

• RCR Category 2: Benign finding

• RCR Category 3: Indeterminate or probably benign finding with a small risk of malignancy such that further investigation is indicated

• RCR Category 4: Findings suspicious of malignancy with a moderate risk of malignancy such that further investigation is indicated

• RCR Category 5: Findings highly suggestive of malignancy with a high risk of malignancy such that further investigation is indicated

9. All three radiologists who gave evidence were in agreement that classification as RCR Category 2 carries a significant implication, in that it indicates that the reporting radiologist has a very high degree of confidence that such findings as have been made are benign, as understood in clinical practice, with the consequence that no recall or further investigation is required at that time.

10. While absolute certainty is neither achievable nor claimed in radiology, RCR Category 2 is reserved for findings that clearly fall below the threshold for further assessment when viewed in their full clinical and radiological context.

Relevant Radiological Concepts

11. It is also necessary to outline certain radiological concepts in order to understand the nature of the dispute between the experts.

12. Microcalcifications are tiny deposits of calcium within breast tissue, usually visible on mammography as small white dots. They are common and most are benign. Their significance depends on number, morphology (shape and form), and distribution (spatial arrangement).

13. A cluster is conventionally described as involving five or more microcalcifications. The significance of a cluster lies in the fact that a grouping of microcalcifications within a confined area may, depending on its characteristics, carry a risk of malignancy.

14. Amorphous microcalcifications are without clear shape, and may appear hazy or powdery. They are frequently benign but carry a risk of malignancy that must be assessed contextually.

15. Pleomorphic microcalcifications vary in size and shape and are more often associated with malignancy.

16. Linearity or linear-branching may suggest ductal involvement and is a feature associated with malignancy.

17. Ductal carcinoma in situ (‘DCIS’) may present as clusters of microcalcifications aligned along ducts.

18. Early malignancy may not be visible on mammography, and microscopic disease may be present long before imaging thresholds are crossed.

Factual Summary

19. On 29 May 2019, the plaintiff, then 45, presented at University Hospital Waterford symptomatic breast clinic with concerns relating to her left breast in the context of a significant family history of breast cancer, including a first-degree relative. Her symptoms were recorded as a ‘discrete breast or axillary lump’ and prompted radiological assessment. A mammogram of both breasts was performed. The report described ‘bilateral patchily glandular breasts without suspicious features or microcalcifications’. The mammogram was assessed as normal and classified as RCR Category 1. There is no criticism by either party of that assessment albeit that the plaintiff’s radiological expert said that he would have classified it as RCR Category 2.

20. In light of her family history, the plaintiff was enrolled in a surveillance programme. Surveillance mammography differs from symptomatic assessment in that imaging is undertaken at regular intervals to detect early changes in individuals known to be at increased risk, even in the absence of new symptoms. The purpose of surveillance is the early detection of radiologically visible disease. It is not designed to detect disease present only at a microscopic level.

21. On 23 October 2020, the plaintiff underwent the first of two surveillance mammograms relevant to this case. The first study was reported by Dr Carmel Ann Daly, consultant radiologist, who compared the mammogram with the preceding study in 2019 and found that glandular involution (replacement of milk-producing glandular tissue with fatty tissue) had occurred in the interval but without any sinister interval change. The mammogram was assessed as benign and classified as RCR Category 2.

22. Dr Daly was called by the defendant to give evidence on this issue. She described and demonstrated to the Court three microcalcifications in the area under scrutiny on the 2020 mammogram, which she characterised as being benign in morphology and as showing, at most, a form of quasi-alignment on one projection falling short of a clinically significant linear pattern. She interpreted the appearances as consistent with background benign microcalcifications that had been made more visible by glandular involution and did not regard them as representing new pathological change. Accordingly, she classified the imaging as benign, with no recall for further assessment.

23. A further surveillance mammogram was performed on 7 January 2022. The study was reported by Dr Laura Sweeney, consultant radiologist, who compared it with the 2019 and 2020 studies and found that there were no suspicious interval changes against a background of scattered microcalcifications in both breasts. She classified the mammogram as RCR Category 2. It was common case that a biopsy, if performed at that time, would likely have revealed cancer or pre-invasive disease. Dr Sweeney was not available to give evidence, through circumstances entirely unconnected to these proceedings. The Court’s assessment is therefore confined to the propriety of the classification made, by reference to the expert evidence adduced, and does not involve any finding as to her personal competence or clinical practice more generally.

24. On 2 May 2023, the plaintiff underwent a third surveillance mammogram. It is agreed that the imaging demonstrated conspicuously malignant features, and that biopsy confirmed invasive ductal carcinoma. The adequacy of that assessment which was carried out by Dr Sweeney is not in issue.

Expert Evidence

25. The core dispute between the expert witnesses does not concern the existence of breast cancer by 2023, nor the biological fact that the disease developed over time but rather, the interpretation of subtle mammographic appearances in 2020 and 2022, and whether those appearances crossed the threshold for recall under accepted professional standards.

26. Expert radiological evidence for the plaintiff was given by Dr Steven Allen, a consultant radiologist at the Royal Marsden Hospital in London, and for the defendant by Dr Sylvia O’Keeffe, a consultant radiologist at St. James Hospital in Dublin. Both experts are experienced consultant radiologists specialising in breast imaging. Both witnesses gave their evidence carefully and in good faith. The divergence between them is therefore not explicable by lack of expertise, but by differences of judgment applied at the margins of radiological interpretation, particularly in relation to microcalcifications.

Points of Agreement

27. Before addressing the areas of disagreement, it is important to identify the substantial body of expert evidence that was not in dispute. Both experts accepted the following propositions:

(1) the plaintiff was asymptomatic in 2020 and 2022 but had a relevant family history of breast cancer;

(2) both mammograms were assessed in the usual way by reviewing both breasts on both the craniocaudal (top-down) and mediolateral oblique

(side) views;

(3) the assessments were carried out within a structured clinical system with established protocols and multidisciplinary oversight;

(4) microcalcifications are a common mammographic finding, particularly in perimenopausal women in their forties, as was the plaintiff;

(5) the interpretation of low-number microcalcifications is among the most challenging tasks in breast radiology;

(6) both intra-observer and inter-observer variability are well recognised and documented in the relevant peer-reviewed literature;

(7) the cancer diagnosed in 2023 likely had a biological origin several years earlier;

(8) mammography has inherent limitations in detecting early or microscopic disease;

(9) classification under the RCR system, particularly at Categories 2 and 3, involves the exercise of professional judgment such that classification under the RCR system is not a purely mechanical exercise.

28. Both experts accepted that hindsight bias must be avoided, and that the task is to assess what should reasonably have been perceived and acted upon at the time, based on the images then available.

Applicable Legal Principles

29. The applicable legal principles governing claims of clinical negligence are well established and derive from the decision of the Supreme Court in Dunne v National Maternity Hospital [1989] IR 91, as clarified and reaffirmed in Morrissey v Health Service Executive [2020] IESC 6.

30. The standard of care is that of an ordinarily competent practitioner exercising the ordinary skill appropriate to the profession. Where the allegation concerns the exercise of professional judgment, rather than the breach of an established practice or mandatory rule, negligence is not established merely because another competent practitioner would have acted differently or adopted a more cautious course. The Court must determine whether the practice relied upon by the defendant is one that is supported by a responsible body of professional opinion, and, if so, whether that opinion withstands logical scrutiny.

31. In cases of radiological interpretation, particularly in relation to mammographic microcalcifications, where one is relying on very small and faint radiological features, the exercise of professional judgment frequently involves unavoidable perceptual difficulty and inter-observer variability. Competent practitioners may therefore differ in their perception and interpretation of subtle features. The existence of such disagreement does not of itself establish negligence.

32. While the Court accepts that radiological interpretation in this area is characterised by inherent perceptual difficulty and interpretative uncertainty, those features do not of themselves immunise a clinician from scrutiny or from liability. The existence of such difficulty informs the evaluation of what is reasonably open to a competent practitioner, but does not displace the requirement that the decision reached must fall within the range of acceptable professional judgment.

33. Where experienced experts differ, that issue is resolved not by preferring one clinical opinion over another, but by examining whether the defendant’s approach discloses a coherent, rational and professionally recognisable methodology that was conscientiously applied. If so, the Court is not entitled to find negligence merely because an alternative and more cautious approach was available. This is particularly so in cases such as the present, where the classification decision involves the application of a confidence threshold within a graded system, rather than the application of a rigid rule admitting of no margin for evaluation.

34. Morrissey emphasises in particular the need to guard against hindsight bias in cases involving threshold-based decision-making in circumstances of inherent interpretive variability. The fact that an adverse outcome subsequently materialised, or that a different approach might, in retrospect, have led to earlier diagnosis, does not of itself establish breach. The Court rejects the submission made on behalf of the plaintiff that the subsequent development of malignancy in the area under consideration demonstrates that the earlier findings could not have been benign. That is an impermissible hindsight analysis. The task of the Court is to assess whether the contemporaneous interpretation fell within a reasonable range of professional judgment, not to infer error from outcome.

35. Applying those principles in the present case, the central question is whether, on a prospective assessment and allowing for recognised perceptual difficulty and variability in the interpretation of mammographic microcalcifications, it was reasonably open to a competent radiologist exercising ordinary care to classify the plaintiff’s mammograms in 2020 and 2022 as RCR Category 2. The issue is not whether a different classification might have been preferable or safer, but whether the classifications made fell within the range of acceptable professional judgment. The existence of a responsible body of opinion supporting a particular approach does not of itself preclude a finding of breach, however, where, on the facts of the case, the

conclusion reached cannot be regarded as one that was reasonable or logically open.

36. While the classification system is informed by recognised descriptors of morphology, distribution and interval change, its application is not mechanical. The evidence establishes that the assessment of such features involves an exercise of clinical judgment in circumstances of inherent perceptual and interpretative uncertainty. In particular, apparent differences in visibility, number or prominence of microcalcifications may arise from projection, tissue overlap and glandular change, and do not necessarily reflect true pathological progression.

Core Areas of Disagreement

37. The dispute between the two experts may be grouped under five headings:

(1) number of microcalcifications;

(2) morphology;

(3) distribution;

(4) interval change, dominance and stability;

(5) the confidence threshold for recall (RCR Category 2 versus Category 3).

Number of Microcalcifications

38. In relation to the 2020 mammogram, Dr Allen identified a cluster of five or six microcalcifications within a confined region measuring approximately 7mm which he maintained were new and not present on the 2019 mammogram. He accepted, however, that counting and measuring at this scale is “somewhat difficult” because fine flecks can “blend in with” or be “masked” by dense background tissue. In relation to the 2022 mammogram, Dr Allen did not specify the number of microcalcifications visible to him but stated that he could see “a few more flecks” which he said were “more conspicuous” and further noted a slight increase in the focal area from 7mm to 8mm. Dr O’Keeffe identified and demonstrated three definite microcalcifications, with a “potential” fourth on the 2020 mammogram and four definite and, on close magnification, a possible fifth calcification on the mediolateral oblique (‘MLO’) view of the 2022 mammogram.

Morphology

39. It is common case that microcalcifications that are round and punctate are benign. Both experts agreed that the relevant microcalcifications visible on both mammograms were not round and punctate, and therefore not definitively benign in their morphology. Dr Allen characterised the demonstrated microcalcifications both from 2020 and 2022 as pleomorphic, meaning that they were variable in size and shape, and therefore often associated with malignancy. Dr O’Keeffe disagreed, contending that pleomorphism presupposes a level of visual clarity which she did not consider to be present. She characterised the microcalcifications as amorphous, describing them as faint, hazy, and of low density, without sharp margins or clear internal structure. She considered that their amorphous appearance, assessed in context, did not preclude a benign classification.

40. Dr Allen accepted that the microcalcifications were subtle and that categorisation with precision is difficult, particularly when dealing with a small cluster. He maintained nonetheless that pleomorphism was the closest classification. Dr O’Keeffe emphasised that amorphous microcalcifications are extremely common in pre-menopausal women and frequently benign. She also accepted, however, that they can be malignant but said that amorphous morphology alone does not mandate recall or biopsy. She maintained that the malignancy risk from amorphous microcalcifications must be assessed by reference to context, including by reference to number, distribution, dominance, bilaterality and interval change.

41. It is significant that Dr Allen did not use the term “pleomorphic” in his initial report for the Court, and that this description emerged at a later stage. While this does not of itself undermine the reliability of his evidence, it reflects a degree of refinement in expression which the Court takes into account in assessing the weight of that aspect of his evidence.

Distribution

42. Both experts agreed that the distribution of microcalcifications is of particular importance, as a linear or ductal pattern may indicate deposition along a duct and is therefore regarded in breast radiology as a feature associated with malignancy.

43. Dr Allen placed reliance on the distribution of the microcalcifications, which he described as linear. He explained that linearity may relate either to the shape of the microcalcifications themselves or to the way they are distributed, and that, in this case, it was the distribution which concerned him. Dr Allen accepted that the appearances did not represent classic linear branching microcalcifications and acknowledged that limitation. Nonetheless, he stated that the alignment of the microcalcifications appeared to him to follow a line consistent with ductal distribution and contributed to his concern. Dr Allen did not regard the alignment as a chance distribution of background microcalcifications. While he accepted that the pattern was not a textbook example of linear branching, he maintained that the linear distribution of the microcalcifications, taken together with their number and other features, supported recall for further assessment.

44. Dr O’Keeffe distinguished between true linear distribution and incidental alignment of a small number of dots. In relation to the 2020 mammogram, she described the three microcalcifications that she saw as “not in a line” but rather, on the MLO projection, as forming “a circle or a triangle”. Dr O’Keeffe did, however, accept that three and possibly four microcalcifications could be seen to lie in a line on the 2022 MLO projection but did not accept that this amounted to true linear or ductal distribution, emphasising the small numbers involved and the absence of consistent appearances across projections.

Interval Change, Dominance and Stability

45. Dr Allen’s evidence was that there was interval change between 2019 and 2020, in that microcalcifications visible in 2020 were new when compared with 2019. He treated the appearance of new microcalcifications in that interval as abnormal and not explicable by background change alone. He regarded the persistence of the same microcalcifications into 2022, with greater conspicuity, as reinforcing concern rather than providing reassurance. In his opinion, the absence of dramatic progression did not imply benignity, particularly given the known behaviour of slowly evolving DCIS, which may show limited or subtle short-term change on mammography.

46. Dr O’Keeffe accepted that there were more microcalcifications visible in 2020 than in 2019 and accepted that two of the microcalcifications seen in 2020 were not appreciable on the 2019 images. She did not, however, regard that difference as suspicious interval change. She explained that in a woman in her forties, amorphous microcalcifications commonly emerge or become newly appreciable due to changes in background tissue, projection and glandular involution, without pathological significance. Between 2020 and 2022, Dr O’Keeffe accepted that the microcalcifications were more conspicuous, particularly on the MLO view, but attributed this to improved visibility rather than to disease progression. She characterised the appearances across 2020 and 2022 as radiologically stable in terms of concern, with no convincing increase in number, no adverse morphological evolution, and no suspicious distributional change.

47. Dr Allen regarded the group of microcalcifications in the upper anterior left breast as a dominant focus, in that it formed a discrete and focal abnormality which, in his view, stood out from the surrounding breast tissue and warranted recall. He did not consider the presence of other microcalcifications elsewhere in the breast to detract from that assessment.

48. Dr O’Keeffe disagreed, stating that dominance must be assessed by reference to the background pattern of microcalcifications within the breast as a whole. She demonstrated that similar small groupings of amorphous microcalcifications were present elsewhere in the breast, particularly on the 2022 images, and considered that the group under scrutiny did not stand out in number, morphology, or appearance, and therefore could not properly be regarded as a dominant feature.

49. In considering the significance of an interval change, the experts also addressed the effect of projection and breast density on the visibility of microcalcifications. Dr O’Keeffe explained that the visibility of microcalcifications is influenced by projection and surrounding breast tissue, and that in women in their forties, changes in breast density, including glandular involution, may render existing microcalcifications more conspicuous without representing pathological progression. On that basis, she attributed the increased conspicuity of the microcalcifications in 2022, particularly on the MLO view, to improved visibility rather than to adverse change. Dr Allen accepted that projection and tissue overlap can, in general, affect the visibility of microcalcifications. He rejected that explanation in this case, stating that he would disagree that projection accounted for the appearances and that the microcalcifications were there to be seen on the earlier images.

50. Both experts accepted that DCIS may evolve slowly, and that lack of dramatic change does not exclude malignancy. Dr Allen stated that low-grade DCIS can change very slowly. In his view, the absence of rapid progression between 2020 and 2022 did not negate malignancy but was consistent with early malignant change. Throughout his evidence, Dr Allen maintained that the appearance of new microcalcifications in 2020, followed by their persistence and increased conspicuity in 2022, supported recall for further assessment.

Confidence Threshold for RCR Category 2

51. The authorities establish that the standard of approach is to be derived from the evidence of the relevant profession.

52. Dr Allen’s evidence was that unless a radiologist can be satisfied with “absolute confidence” that microcalcifications are benign, recall is mandated. He rejected the proposition that residual uncertainty or doubt could appropriately be accommodated within RCR Category 2, stating that if a radiologist cannot say that findings are definitively benign, recall for further assessment is required. Dr Allen rejected the suggestion that the threshold for recall involves a discretionary judgment balancing competing considerations and rejected the relevance to individual classification decisions of broader population-based considerations such as recall rates, resource utilisation, and the avoidance of unnecessary biopsies. Applying that approach, Dr Allen maintained that the mammographic appearances in 2020 and 2022 could not be said to be benign with absolute confidence and therefore did not meet the threshold for classification as RCR Category 2.

53. Dr O’Keeffe, while not directly addressing the threshold of confidence required for Category 2 classification, stated that “if you have any doubt that it looks atypical, you call it a (RCR Category) 3”. Read as a whole, however, Dr O’Keeffe’s evidence was more nuanced. She explained that amorphous microcalcifications are common, particularly in women under 50, as was the plaintiff, that the vast majority are benign, and that biopsy is invasive, resource-intensive, and potentially harmful, as well as burdensome and distressing for patients. She said that not all possibly malignant microcalcifications can or should be biopsied so that a judgment call must be made. Dr O’Keeffe also distinguished between different models of clinical practice. She accepted that, unlike her own consensus practice, in which she could consult colleagues where there was uncertainty as to whether a mammogram met the threshold for recall, Dr Daly and Dr Sweeney worked in a ‘single read’ practice without the benefit of routine collegial review. She said that in such a setting, one must “operate on a lower threshold and biopsy more”, with the result that if there is “any doubt, they have to biopsy”.

54. The Court does not accept that the expressions ‘absolute confidence’ and ‘absence of any doubt’ can be treated as identical or interchangeable formulations. Both formulations are merely descriptive expressions used by the relevant witnesses and do not represent a legal standard. The applicable standard remains that derived from professional practice and assessed in accordance with Dunne and Morrissey. While there is overlap in the level of confidence sought to be conveyed, the evidence as a whole demonstrates that both were used in a non-literal sense within a context of acknowledged uncertainty and the need for evaluative judgment. Properly understood, in a domain where professional confidence must operate within conditions of unavoidable interpretative difficulty and clinical uncertainty, both witnesses were describing a confidence threshold for RCR Category 2 that is very high, well above mere likelihood, but short of absolute or mathematical certainty. What is required is the highest level of professional confidence realistically attainable within a domain of acknowledged interpretative difficulty and clinical uncertainty where some possibility of error necessarily remains for clinical evaluation and judgment. This synthesis reflects the Court’s assessment of the applicable standard of approach rather than reflecting a complete convergence of the terminology used by the experts. The Court approaches their evidence on that basis and applies the legal test in Dunne accordingly.

55. The plaintiff’s legal submissions placed considerable reliance on the formulation of the standard of approach in cases such as Morrissey v Health Service Executive and Penney and Ors v East Kent Health Authority [2000] Lloyd’s Rep Med 41, in which it was stated that a screener should not classify a sample from a cervical smear as normal where there is “any doubt”. Those authorities, however, arose in the different context of cytological screening, where the evidence established a particular operational rule governing binary classification decision-making in that domain.

56. The present case, by contrast, concerns radiological interpretation within a graded classification system requiring a judgment by way of evaluation as to whether findings may properly be regarded as benign to a sufficiently high level of professional confidence. The expert evidence in this case does not establish a uniform professional rule that the presence of any degree of doubt, however slight, necessarily precludes classification as RCR Category 2. While both experts emphasise that a very high level of confidence is required, the evidence also reflects the reality that the assessment of subtle microcalcifications involves the exercise of professional judgment in circumstances of acknowledged inherent interpretative uncertainty and variability.

57. In those circumstances, the Court is satisfied that the applicable standard of approach is that RCR Category 2 is reserved for findings in respect of which a radiologist can properly maintain the requisite level of confidence for benign classification as outlined above and that the existence and degree of any residual doubt is to be assessed within the exercise of that professional judgment in its particular evidential context.

Use of Medical Literature

58. Both experts relied on a body of peer-reviewed literature addressing the interpretation of mammographic microcalcifications, reporting classification systems, recall thresholds, and inter-observer variability. The literature was not relied upon to establish novel scientific propositions, but to support differing views as to how borderline appearances should reasonably be assessed in clinical practice.

59. Dr Allen relied on the literature to emphasise that RCR Category 2 is reserved for findings that are clearly benign, and that where microcalcifications are not unequivocally benign, on assessment of a mammogram, recall is required. He accepted that the literature recognises interpretive difficulty and variability but maintained that these considerations do not justify non-recall where abnormal features are present.

60. Dr O’Keeffe relied on the same literature to emphasise that amorphous and lownumber microcalcifications are extremely common, particularly in women in their forties, and that the majority are benign. She pointed to published data documenting substantial inter-observer variability in the assessment of microcalcifications, and to literature demonstrating that lower recall thresholds increase benign biopsy rates without proportionate clinical benefit. She relied on these sources to support the proposition that non-recall in low-suspicion cases falls within the range of accepted professional practice.

61. The experts did not materially disagree as to the content of the literature. Their disagreement lay in the weight to be attributed to it, and in how its implications were to be applied to the specific mammographic appearances under review.

Variability and the Significance of Disagreement

62. Both experts accepted that radiologists may disagree, even when exercising due care, particularly in relation to microcalcifications.

63. Dr Allen maintained that this was not a case where reasonable disagreement could arise. Dr O’Keeffe considered it a paradigm case of borderline interpretation within accepted limits.

64. It is not without relevance that Dr Allen accepted, in cross-examination, that in another case in which he had advised as an expert, he had reached differing conclusions on the same mammogram, illustrating the reality of interpretive variability. The Court treats that evidence as illustrative of the broader issue of interpretative variability in this field, rather than as a reflection on the reliability of his evidence in the present case. The Court has considered this, together with other matters raised in cross-examination concerning the development of aspects of his description of the relevant findings over time.

Findings – 2020 Mammogram

65. The Court accepts that the plaintiff’s case depends on the cumulative effect of multiple features, including interval change, number, morphology, distribution and clinical context, including family history. The issue is whether, at the relevant time, those features when taken together, deprived a competent radiologist of a reasonable basis for maintaining a very high level of confidence in benignity.

66. The Court accepts that Dr Daly’s approach to the review of the 2020 mammogram was methodical, professionally recognised and consistent with accepted radiological practice. The mammographic appearances in 2020 were characterised by a very small number of clearly visible microcalcifications, variously identified as three and possibly four within a confined area. The plaintiff’s expert identified additional fainter foci which he would have counted as microcalcifications, thereby potentially exceeding the commonly referenced numerical threshold for a cluster. However, both experts accepted that precise counting at this scale is inherently difficult and subject to perceptual variability. Importantly, the proposition that more than a definite three and possibly four microcalcifications were present depended upon the inclusion of additional radiological features identified as microcalcifications seen by Dr Allen but not Dr O’Keeffe.

67. In terms of morphology, the microcalcifications demonstrated in 2020 were accepted to be irregular but subtle and difficult to categorise. Dr Allen regarded pleomorphism as the most appropriate descriptor, whereas Dr O’Keeffe characterised the appearances as amorphous. Both experts accepted that amorphous microcalcifications are common, particularly in women in their forties, and are frequently benign, albeit not invariably so. The Court is satisfied that, at that stage, the morphology did not preclude a reasonable radiologist from perceiving the microcalcifications to be amorphous and benign in their particular clinical context.

68. The distribution of the microcalcifications in 2020 did not demonstrate a classic linear or ductal pattern. The magnified image of the relevant area shown by Dr O’Keeffe demonstrated, at most, two microcalcifications that were “close to each other” but “not in a line”. While Dr Allen expressed concern regarding alignment of what he perceived to be a cluster of microcalcifications, he accepted that the appearances did not amount to textbook linear branching microcalcifications. The Court accepts Dr O’Keeffe’s evidence that incidental alignment of a very small number of microcalcifications does not of itself constitute a suspicious distribution, particularly where correlation between projections was limited.

69. The Court has also considered the significance of interval change between 2019 and 2020. It is accepted that some microcalcifications visible in 2020 were not appreciable on the 2019 study. However, the evidence establishes that, in women of the plaintiff’s age, changes in breast density, glandular involution, and projection may render microcalcifications newly visible without pathological significance.

70. The Court is satisfied that, in 2020, the appearance of a small number of amorphous microcalcifications against that background did not mandate recall as a matter of professional obligation.

71. Crucially, and in contrast to 2022, the 2020 mammogram represented the first presentation of subtle microcalcifications without demonstrated persistence, without convincing dominance against the background of bilateral microcalcifications, and without a cumulative combination of features sufficient to undermine the level of confidence required for RCR Category 2. At that stage, the case lay within an area of acknowledged interpretive difficulty in which reasonable and competent radiologists could differ as to whether recall was warranted. That the same appearances were subsequently found to be associated with malignancy does not retrospectively alter that assessment. The Court accepts that, with the benefit of hindsight, the microcalcifications identified in 2020 and 2022 were likely manifestations of the cancer diagnosed in 2023. However, that retrospective fact does not determine breach. The issue is whether those findings, as they appeared at the time and without knowledge of outcome, required recall under accepted professional standards. The evidence establishes that what were early malignant microcalcifications in 2020 were indistinguishable from benign findings and did not reach the threshold for intervention.

72. Dr Allen’s evidence reflects that a more cautious approach was available in 2020, and the Court accepts that such an approach might, with hindsight, have resulted in earlier diagnosis. However, under the principles stated in Dunne v National Maternity Hospital and Morrissey v Health Service Executive, the existence of an alternative more cautious articulation of clinical practice does not establish the minimum standard required or breach of that standard unless the approach actually adopted falls outside the range of acceptable professional practice. Even if the Court were to prefer the interpretation advanced by Dr Allen, it would remain satisfied that the classification adopted fell within the range of reasonable professional judgment.

73. Having regard to the low number of microcalcifications, their subtle and commonly encountered amorphous appearance in women of the plaintiff’s age, the absence of a convincing suspicious distribution, and the availability at that time of a plausible benign explanation grounded in glandular involution and projectional effects, together with recognised variability in interpretation at this level, the Court is satisfied that classification of the 2020 mammogram as RCR Category 2 remained within the range of acceptable professional judgment.

74. While the Court accepts that the features identified by the plaintiff’s expert could give rise to concern, it is not satisfied that, in 2020, those features had coalesced into an overall presentation sufficient to deprive a competent radiologist of a reasonable basis for classifying the findings as benign. At that stage, the case lay within a recognised area of interpretive difficulty in which reasonable and competent practitioners could differ.

75. Accordingly, the plaintiff has not discharged the burden of establishing breach of duty in respect of the October 2020 mammogram.

Findings – 2022 Mammogram

76. The defendant submits that the differences between the 2020 and 2022 appearances are attributable to projectional and perceptual factors rather than to any true interval change. The Court does not accept that the differences can be explained solely on that basis. While such factors are relevant, they do not adequately explain the persistence, focality and increased conspicuity of the features present by 2022.

77. Taken in isolation, each of the individual features identified on the 2022 mammogram, namely, the limited number of microcalcifications, their plausibly amorphous morphology, the absence of classic linear branching, and the absence of dramatic interval progression, could, in a different clinical context, fall within an area of legitimate professional debate. The existence of those features does not of itself establish that non-recall was impermissible.

78. However, the Court is not required to consider those features in isolation. By January 2022, there existed a cumulative constellation of features which materially altered the evaluative context. These included the persistence over time of the same focal grouping, the presence of what was accepted by the defendant’s expert to be at least four, and possibly five, microcalcifications which were, at a minimum, amorphous in their morphology within a confined area, and the fact, not in dispute, that three and possibly four of those microcalcifications were aligned on the MLO projection, although that alignment did not amount to a classical ductal distribution.

79. In that regard, the Court attaches particular significance to the progression, on the defendant’s case, in both number and distribution of microcalcifications as between 2020 and 2022. In 2020, the finding consisted of no more than three, and possibly four microcalcifications, with what appeared in Dr O’Keeffe’s demonstration to the Court to be no more than two capable of being viewed as merely “close to each other”. By 2022, this had developed to at least four definite microcalcifications, with a possible fifth, potentially constituting a cluster within a confined area, with three and possibly four, demonstrating alignment on the MLO projection, albeit without consistent correlation across projections. This represented not merely an incremental increase, but a qualitative change in the character of the finding, from a small number of possibly incidental microcalcifications to a more coherent focal grouping exhibiting persistence together with elements of potential clustering and linearity. The significance of these cumulative features lies not merely in their number or description, but in their combined effect on the sustainability of the high level of professional confidence in benignity required for RCR Category 2 classification.

80. Critically, these features fell to be assessed against the background that the appearances could no longer reasonably be regarded as merely transient or newly emergent background change. Whatever uncertainty reasonably existed in 2020, the persistence and progression of the appearances into 2022 materially altered their significance. In that context, the submission that the appearances under consideration formed part of a wider pattern of similar microcalcifications elsewhere in the breast does not materially advance the defendant’s case. The Court accepts that the presence of background microcalcifications is a relevant consideration, particularly in deciding whether a finding is dominant or exceptional. However, the proper question is whether the group under consideration possessed distinguishing features sufficient to warrant separate evaluation. By 2022, the Court is satisfied that the focal grouping at issue exhibited such features. While the presence of bilateral and background microcalcifications is accepted, the evidence does not establish that the other groupings relied upon were equivalent in their relevant characteristics. In particular, the Court is satisfied that the grouping of microcalcifications under scrutiny differed in its persistence, focality and development both in number and alignment such that it had crossed a qualitative threshold.

81. The Court accepts that benign amorphous microcalcifications are common in women in their forties and that projectional differences, breast density, and interpretative variability may, in other cases, support a conclusion of stability. However, in the present case, the explanatory force of those contextual factors, when confronted with the cumulative features present by January 2022, was no longer sufficient to sustain a RCR Category 2 classification. The issue is not whether malignant or suspicious features were demonstrated, nor whether biopsy was inevitable, but whether it remained professionally tenable to make a positive and very high confidence finding of benignity.

82. Applying the principles articulated in Dunne v National Maternity Hospital and Morrissey v Health Service Executive, the Court is satisfied that, by this point, the maintenance of an RCR Category 2 classification required a level of confidence in benignity that could no longer reasonably be sustained within the range of professional judgment reasonably open on the evidence. As previously outlined in the Court’s analysis above, Category 2 requires not merely the absence of features mandating recall, but the positive ability to characterise the findings as benign to a very high degree of professional confidence.

83. In reaching that conclusion, the Court emphasises that it is not rejecting either the evidence or the general approach of Dr O’Keeffe, which the Court accepts to be coherent, rational, and reflective of accepted professional practice. Nor is the Court substituting its own clinical judgment for that of a specialist radiologist. Rather, the Court’s function is to determine whether, on the facts of this particular case, and having regard to the cumulative constellation of features present by January 2022, the

conclusion actually reached remained one reasonably open within the framework of that accepted methodology. The difficulty for the defendant lies not in any deficiency in the general approach described, but in the application of that approach to a factual scenario in which the persistence, focality and developing number and alignment of microcalcifications had, by 2022, materially undermined the basis upon which a benign classification could confidently rest. It is in that limited but critical respect that the Court finds the boundary of permissible professional judgment to have been crossed.

84. The Court accepts the evidence of Dr O’Keeffe that a number, and possibly a majority, of competent radiologists might have classified the mammogram as RCR Category 2. However, the standard articulated in Dunne and reaffirmed in Morrissey does not turn on the numerical weight of professional opinion, but on whether the

conclusion reached was one that remained reasonably open to a competent practitioner exercising ordinary care. A court is not absolved from scrutiny merely because a body of opinion exists, still less because it is said to be widely held within the profession. The question is whether the conclusion reflects a coherent and rational application of the relevant clinical methodology to the facts of the case.

85. The distinction between RCR Category 2 and Category 3 is of particular importance in this context. Category 2 is reserved for findings that can properly be characterised as benign to a very high level of professional confidence. Category 3, by contrast, encompasses findings which, while not overtly suspicious, cannot be characterised to that level of confidence and are therefore regarded as ‘indeterminate or probably benign’, requiring further investigation. By January 2022, the justification advanced for maintaining a Category 2 classification rested not on the positive identification of benign features to the level of professional confidence identified above, but rather on the absence of sufficiently suspicious features to mandate recall. In the Court’s view, that reasoning, taken at its highest, supports a conclusion of probable, but not definitive, benignity, but does not sustain the very high degree of confidence required for a Category 2 classification. In those circumstances, the appropriate classification, even within the framework of the defendant’s own approach, was RCR Category 3.

86. The issue is not whether recall would have been preferable, but whether, within the accepted methodology, a classification of benignity remained logically defensible. At that stage, it was no longer reasonably open to a competent radiologist, exercising ordinary care, to maintain the requisite very high level of confidence that the appearances were benign, and the findings should, at minimum, have been regarded as indeterminate, thereby requiring recall for further assessment.

Decision

87. Accordingly, the Court concludes that, by January 2022, classification of the mammogram as benign fell outside the range of acceptable professional judgment, and that the failure to recall the plaintiff for further assessment at that time constituted a breach of duty.