Impact of skin biopsy practices: A comprehensive nationwide study on skin cancer and melanoma biopsies

Abstract

Background

Due to a multitude of factors, skin cancer incidence is increasing and challenges medical professionals in biopsy decision‐making. While skin cancer may have a profound impact on the patient and be costly for society, there is little knowledge about the number and cost of benign skin lesions biopsied as collateral damage.

Objectives

This study evaluates the number and costs of skin biopsies in Denmark over 15 years, focusing on benign and malignant skin lesions and melanomas across medical settings. It aims to determine the benign to malignant ratio (BMR) and number needed to biopsy (NNB) and estimate the direct cost of benign skin lesion biopsies in the Cancer Pathway from the perspective of a public healthcare system.

Methods

The study included 4,481,207 biopsy specimens from January 2007 to June 2022 from the Danish Pathology Data Bank, of which 151,988 from the Cancer Pathway were included in the primary analysis of BMR. The national reimbursement rates for biopsies were used, alongside histopathological examination costs extracted from several pathology departments, for a Monte‐Carlo simulation of a simple cost and sensitivity analysis.

Results

The number of biopsies increased by 39.1% from 2007 to 2021. Overall BMR for malignancy was 4.1:1, and NNB for melanoma was 31.8, but biopsies performed on clinical suspicion of malignancy or melanoma had a BMR and NNB of 1.5:1 and 2.8, respectively. The cost of benign skin biopsies performed on suspicion of cancer or melanoma in 2021 was €6.6M, predominantly in hospitals.

Conclusions

A healthcare system that employs filtering functions before biopsy of skin lesions can achieve some of the lowest BMR reported in the world, but with most benign skin lesion excisions due to suspicion of malignancy performed in the expensive hospital setting. Including clinical reason for biopsy in diagnostic accuracy studies using NNB is crucial.

Skin biopsy decision‐making is challenging. From a comprehensive national pathology registry including all biopsies, we investigated benign to malignant ratios (BMR) and number needed to biopsy (NNB) for lesions that might clinically mimic melanoma, for all biopsies and biopsies clinically suspected of cancer or melanoma. From 4.5 million included biopsies we found that (1) clinical suspicion of cancer impacts the BMR for malignancy and drastically affects the NNB for melanoma. (2) The amount of skin biopsies increased by 39% from 2007 to 2021. (3) The public funded healthcare system in Denmark spends €6.5M on benign skin lesions suspected of being melanoma—with the vast majority of these expenses from the hospital sector of the healthcare system. We therefore conclude that (1) layered triage helps separate low‐risk from high‐risk biopsies to achieve world‐class NNB for melanoma and (2) it is crucial to include the clinical reason for biopsy in diagnostic accuracy studies focusing on NNB.

Why was the study undertaken?

This study aims to investigate how a layered healthcare system affects the biopsy rates in the separate layers, and pinpoint areas of potential improvement in terms of biopsy rates and healthcare costs, using comprehensive Danish registry data.

What does this study add?

This study presents values of benign to malignant (BMR) and number needed to biopsy (NNB) both overall but also specifically for biopsies presumably performed due to a suspicion of cancer or melanoma in various medical specialties and sectors, including the cost of benign biopsies of lesions suspected of melanoma.

What are the implications of this study for disease understanding and/or clinical care?

NNB varies depending on the clinical suspicion, which should be accounted for in diagnostic accuracy studies using NNB. Layered filtering of patients before biopsy effectively lowers the BMR and NNB, but such a healthcare setup is expensive.

INTRODUCTION

Skin cancer and melanoma incidences have been increasing for many years, likely caused by a multitude of factors, including UV radiation, an ageing population and potential overdiagnosis. ¹ , ² , ³ It has been stipulated that mere clinical experience does not increase the diagnostic accuracy of primary care physicians (PCPs). ⁴ , ⁵ The clinical uncertainty associated with perceived diagnostic inaccuracy leads to defensive medical practice, for example, a lower threshold for referral and biopsies. ⁶ A meta‐analysis by Petty et al. shows that the number needed to biopsy (NNB) to find one case of melanoma varies significantly among clinical settings, with an overall of 9.7 ranging from 5.9 among skin cancer specialists to 22.6 among PCPs. ⁷ Heightened awareness and too many biopsies is thought to be leading causes of overdiagnosis of melanoma for pathologists, ⁶ , ⁸ , ⁹ , ¹⁰ in part because the histopathological evaluation of melanoma, just like the clinical evaluation, may be challenging. ¹¹ , ¹² , ¹³

Skin cancer, including melanoma, is a costly group of diseases due to high incidence, sometimes expensive treatment modalities and potentially fatal outcome. ² , ³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ However, there is little knowledge about the amount and cost of the benign skin lesions treated as collateral damage when diagnosing malignant skin tumours.

The aim of this study was to evaluate the number of skin biopsies performed among healthcare professionals in different healthcare settings over 15 years and measure the number of benign and malignant skin lesions and melanomas in various medical settings in terms of benign to malignant ratio (BMR) and NNB, focusing on the three main medical specialities involved in triage and treatment of skin cancer and melanoma; primary care, dermatology and plastic surgery.

Additionally, a simple cost analysis and estimation of the direct cost of biopsies and histopathologic examination of benign skin lesions in different medical settings from the perspective of a public healthcare system was performed.

MATERIALS AND METHODS

Information about skin specimens referred to pathologists between 1 January 2007 and 30 June 2022 was obtained from the Danish Pathology Data Bank ¹⁹ ; a national registry that holds complete information on all pathoanatomical examinations in Denmark, including topographical location, material type and diagnosis. It also includes the requisition priority, a Pathway, chosen by the doctor performing the biopsy:

• Urgent: Examinations related to acute treatment‐requiring or life‐threatening conditions.

• Cancer: Examinations that are part of a cancer fast‐track pathway, which for skin tumours are lesions suspected of melanoma. 90% of results expected within six working days.

• Routine: Examinations other than Urgent and Cancer Pathway, for example, inflammatory skin lesions, non‐melanocytic skin cancer and skin lesions removed without suspicion of malignancy., as all skin specimens are to be evaluated by a pathologist. These lesions are processed chronologically. ²⁰ , ²¹ , ²²

Information about healthcare units was obtained from the Danish Healthcare System Organization Registry and the Hospital Department Classification Registry hosted by the Danish Health Data Authority. ²³ , ²⁴

The medical specialities were categorized as PCP, dermatology, plastic surgery and all others and further divided into either primary care, private practice and hospital setting.

Data processing

Skin specimens obtained with excision, incision, curettage, shaving or punch biopsies were categorized in severity classes, using the Systematized Nomenclature of Medicine (SNOMED) code for their histopathological diagnosis, as malignant, in situ, suspected malignant, unknown whether malignant or benign, benign or no diagnosis. ²⁵ Actinic keratosis (code M72850) was classified as benign. Specimens with codes P30624 (re‐resection) and M09451 (no tumour remains), indicating that the lesion was previously biopsied, were excluded. If the histopathological diagnosis was revised, the revision was matched to the original biopsy using a topographical near‐match algorithm (see Supplementary Materials) and the revised diagnosis was used. For further details, see Figure 1.

FIGURE 1.

Flowchart of the in‐ and exclusion process and grouping for the various analyses. Benign melanoma mimics include all subtypes of nevi, lentigo, pigmentation, seborrheic keratosis, haemangioma, dermatofibroma, haemorrhage and hematoma, purpura, verruca for acral lesions and dystrophia and dermatophytosis for lesions on nails.

Benign to malignant ratio

For the BMR analysis, biopsies with no diagnosis were excluded. Biopsies diagnosed as benign were divided by all biopsies diagnosed as malignant, in situ, suspected malignant or unknown whether malignant or benign. Separate analyses were performed for the entire dataset and for lesions referred in the Cancer Pathway, assumably due to a suspicion of malignancy.

Number needed to biopsy

NNB was calculated using the total number of biopsies that might clinically resemble melanoma, including benign lesions that might resemble melanoma referred to as benign melanoma mimics, divided by the number of actual melanomas (including all subtypes of melanoma and melanoma in situ, including lesions histopathologically diagnosed as ‘suspected melanoma’). Benign melanoma mimics included all subtypes of nevi, lentigo, pigmentation, seborrheic keratosis, haemangioma, dermatofibroma, haemorrhage and hematoma, purpura, verruca for acral lesions and dystrophia and dermatophytosis for lesions on nails. See Table S1 for the full list. Separate analyses were performed for the entire dataset and for biopsies in the Cancer Pathway assumably performed due to suspicion of cancer.

Simple cost analysis

The healthcare reimbursement rate for performing a punch biopsy, other biopsy procedures, teledermatology or follow‐up in the different healthcare settings in 2021 was found in the reimbursement agreement for PCPs, ²⁶ private practising dermatologists and plastic surgeons ²⁷ and the Diagnosis Related Group national average expenditures for patients suspected of having melanoma. ²⁸ National healthcare operational costs were found on the regional government website. ²⁹

The economical estimates, including reimbursement for informing the patient of the result and appointments for suture removal, were made without potential expenditures for additional follow‐up surgery/procedures or complications, focusing on the resources spent on potentially avoidable biopsies of benign lesions.

The cost of histopathological examination of benign skin lesions referred in the Cancer Pathway in 2021 was extracted from all 15 pathology departments covering costs of 64.1% of specimens. When the cost was unknown, it was estimated based on the cost of similar specimens.

Economic sensitivity analysis

To understand the potential financial impacts of different clinical approaches to the benign skin lesions in Cancer Pathway compared to current practice, we conducted an economic sensitivity analysis, examining how costs fluctuate under four hypothetical patient management scenarios. Despite current literature, ³⁰ , ³¹ we chose a conservative estimate of 75% specificity when using digital sequential dermoscopic follow‐up to dismiss benign skin lesions, excising the rest. Scenarios 1, 3 and 4 assume that 25% of the patients are seen by a PCP before referral to a plastic surgeon for excision, with the remaining 75% of patients seen by both a PCP and a dermatologist before referral. For all scenarios, incl. current practice, we assume that full excision results in suture removal at the PCP office and biopsy results are delivered via electronic mail from the setting where the excision was performed.

• Scenario 1: Rather than a biopsy, the patient is seen for a sequential follow‐up consultation in the setting where the biopsy was to be performed, after which we assume that 75% would be dismissed, the remaining 25% excised.

• Scenario 2: The patient was never referred but had the lesion excised on the initial PCP visit.

• Scenario 3: The patient is referred from the PCP to a private dermatologist and seen for a follow‐up consultation using digital sequential dermoscopy. After that, 25% are referred to a private plastic surgeon for excision, of which 10% are subsequently referred to a plastic surgery department due to special needs (e.g. large facial lesions) or co‐morbidity.

• Scenario 4: The patient is followed using digital sequential teledermoscopic follow‐up via the PCP. After that, 25% are referred to a private plastic surgeon for excision, of which 10% are subsequently referred to a plastic surgery department due to special needs special needs (e.g. large facial lesions) or co‐morbidity.

Statistical analyses

The data were cleaned, processed and analysed using R ³² in RStudio. ³³

Due to a skewed cost distribution of histopathological examination, several distribution models were tested for best fit and a logarithmical normal distribution was used to estimate the histopathological cost of biopsies with unknown costs. A Monte Carlo simulation with 10,000 iterations introduced variability in the simple cost and financial sensitivity analysis to calculate a realistic range of the actual costs.

For clarity in interpretation when analysing change over time, the two most recent 5‐year periods (January 2012 to June 2017 and July 2017 to June 2022) were isolated for a Z‐test, with the z‐score indicating how many standard deviations the mean of the most recent period was above or below the previous period mean.

Ethical approval

This study was accepted by the Regional Ethics Committee (R‐22040549) and Legal Department of Scientific Research (2022‐495) at the Capital Region of Denmark.

RESULTS

Data overview

The Danish Pathology Data Bank data were cleaned according to Figure 1, resulting in 4,481,207 specimens for further analysis. Overall, the total number of skin specimens in all pathways increased by 39.1% over the study period.

Benign to malignant ratio analysis

The BMR analysis show some variations across settings as shown in Table 1, with an overall BMR of 4.1:1. For the analysis of BMR of biopsies referred using the Cancer Pathway, 151,988 skin biopsies with a diagnosis were included, and showed only slight variations across settings, as shown in Table 2, with an overall ratio of 1.5:1.

TABLE 1.

Distribution of malignancy, melanoma and benign melanoma mimics among all specimens biopsied.

Setting	Benign	Unknown if benign or malignant	Suspected malignant	In situ	Malignant	BMR	Benign MM mimic	Susp. MM	MM	NNB
Primary care	1,115,510	831	1011	8429	50,232		801,770	278	6075	127.2
Dermatology, private	1,396,923	2033	7430	37,473	441,398	2.9:1	800,402	623	14,532	53.8
Dermatology, hospital	174,079	471	1365	8855	54,730	2.7:1	37,366	72	2502	15.5
Plastic surgery, private	331,882	280	261	3312	34,838	8.6:1	237,330	53	2017	115.7
Plastic surgery, hospital	167,121	1105	708	17,300	154,192	1:1	85,263	310	38,328	3.2
All others, private	281,417	578	235	1982	16,917	14.3:1	99,871	25	884	110.9
All others, hospital	134,955	524	251	3804	28,775	4:1	23,823	38	2049	12.4
Overall	3,601,887	5822	11,261	81,155	781,082	4.1:1	2,085,825	1399	66,387	31.8

Note: Gray columns indicate calculated results based on the included columns in the left. BMR: benign to malignant ratio with malignant including in situ, suspected malignant and unknown if benign or malignant; MM: melanoma including in situ melanoma; NNB: number needed to biopsy, calculated as benign melanoma mimics biopsied per MM including suspected melanoma. Benign melanoma mimics include all subtypes of nevi, lentigo, pigmentation, seborrheic keratosis, haemangioma, dermatofibroma, haemorrhage and hematoma, purpura, verruca for acral lesions and dystrophia and dermatophytosis for lesions on nails.

TABLE 2.

Distribution of malignancy, melanoma and benign melanoma mimics among specimens referred in the Cancer Pathway.

Setting	Benign	Unknown if benign or malignant	Suspected malignant	In situ	Malignant	BMR	Benign MM mimic	Susp. MM	MM	NNB
Primary care	1272	3	10	45	415	2.7:1	911	3	108	9.2
Dermatology, private	16,921	66	202	1313	5199	2.5:1	12,864	42	2444	6.2
Dermatology, hospital	6375	29	73	500	2172	2.3:1	3506	16	831	5.1
Plastic surgery, private	2450	9	6	136	698	2.9:1	2045	3	440	5.6
Plastic surgery, hospital	51,419	311	129	7364	32,856	1.3:1	40,223	186	29,076	2.4
All others, private	1446	8	21	82	780	1.6:1	500	2	52	10.3
All others, hospital	12,785	116	88	1082	6678	1.6:1	2680	16	1280	3.1
Overall	92,668	542	529	10,522	48,798	1.5:1	62,729	268	34,231	2.8

Note: Gray columns indicate calculated results based on the included columns in the left. BMR: benign to malignant ratio with malignant including in situ, suspected malignant and unknown if benign or malignant; MM: melanoma including in situ melanoma; NNB: number needed to biopsy, calculated as benign melanoma mimics biopsied per MM including suspected melanoma. Benign melanoma mimics include all subtypes of nevi, lentigo, pigmentation, seborrheic keratosis, haemangioma, dermatofibroma, haemorrhage and hematoma, purpura, verruca for acral lesions and dystrophia and dermatophytosis for lesions on nails.

Most skin specimens referred using the Cancer Pathway were in accordance to guidelines biopsied by hospital plastic surgeons, with relatively few biopsies from PCPs (Figure 2). The z‐test revealed a significant increase in the proportion of malignant cases from 2012–2017 to 2017–2022, with a positive z‐score of 2.8 (p < 0.01) (not shown). Increases were observed among dermatologists, both private (z = 3.4, p < 0.001) and in the hospital (z = 2.0, p = 0.04). Melanoma accounted for slightly more than half of all malignant lesions in the Cancer Pathway (Figure 3).

FIGURE 2.

The distribution of malignancy classes for all lesions sent in Cancer Pathway (stacks) and percentage of biopsies with malignancy (line using secondary y‐axis). Before 2020, clinically suspected melanomas were by Danish Medical Authority guidelines excised only by hospital plastic surgeons. Z: z‐score, positive indicating an increase in the proportion of malignancy; p: p‐values for the difference in the proportion of malignant and benign skin lesions in the time periods indicated by dark and light grey.

FIGURE 3.

Distribution of the malignant diagnoses in skin biopsies submitted using the Cancer Pathway with percentage shown with dotted lines (using secondary y‐axis). BCC: basal cell carcinoma; PCC: squamous cell carcinoma; MM: melanoma.

Number needed to biopsy

The NNB analysis shows drastic variation across settings from 127.2 by PCPs to 3.2 for hospital plastic surgeons, with an overall NNB of 31.8 (Table 1).

For the analysis of NNB of specimens presumably biopsied due to suspicion of melanoma and hence referred using the Cancer Pathway, 96,917 skin specimens diagnosed as melanoma or benign melanoma mimics were included. A nationwide NNB of 2.8 was observed, with further details on variations between settings shown in Table 2.

The number of biopsies increased rapidly across the study period, particularly for plastic surgeons in the hospitals, where most lesions suspected of melanoma are biopsied due to national guidelines ³⁴ , ³⁵ (Figure 4), and 42.2% of these lesions were melanoma.

FIGURE 4.

Biopsies in Cancer Pathway with melanoma and benign melanoma mimics (boxes) and the percentages of melanoma (line using secondary y‐axis). Benign melanoma mimics include all subtypes of nevi, lentigo, pigmentation, seborrheic keratosis, haemangioma, dermatofibroma, haemorrhage and hematoma, purpura, verruca for acral lesions and dystrophia and dermatophytosis for lesions on nails. Before 2020, clinically suspected melanomas were by Danish Medical Authority guidelines excised only by hospital plastic surgeons. Gray shade areas indicate 5 year periods for comparison of malignancy‐proportions using Z‐test. Z: z‐score, positive indicating an increase in the proportion of malignancy; p: p‐values for the difference in the proportion of malignant and benign skin lesions in the periods indicated by dark grey and grey.

Since the implementation of the Cancer Pathway in 2008, the proportion of melanomas has been steady. The z‐test indicated no significant change in the proportion of melanoma between the two periods overall (z = −0.8, p = 0.44), but with some variation among the settings. Analysis of NNB when disregarding Pathways is shown in Figure S1.

Missed melanomas

For the analysis of melanomas in the Routine Pathway, 3,880,190 specimens were included. Overall, 0.63% of lesions submitted in Routine Pathway were melanoma, equivalent to one case of melanoma in every 161 specimens referred. Plastic surgeons in hospitals stand at odds with the rest, with 3.08% of specimens in Routine Pathway showing melanoma (Table 3), worst at the beginning of the period, as displayed in Figure S2.

TABLE 3.

Melanoma specimens referred for histopathological evaluation in Routine Pathway.

Medical specialty	Healthcare setting	Number of melanomas	Total biopsies in Routine pathway	Proportion of melanoma (%)
Primary care physician	Primary care	5529	1,119,784	0.49
Dermatologist	Private practice	8065	1,625,260	0.50
Dermatologist	Hospital	1230	205,391	0.60
Plastic surgery	Private practice	1281	310,440	0.41
Plastic surgery	Hospital	7144	232,321	3.08
All others	Private practice	602	249,698	0.24
All others	Hospital	600	137,296	0.44
All	All	24,451	3,880,190	0.63

Note: Gray columns indicate calculated results based on the included columns in the left.

Simple cost analysis

Of the 9083 benign specimens in the Cancer Pathway in 2021, the total amount spent in 2021 was €6.6 million (95% UI 6.2–7.5), corresponding to 0.04% of the national healthcare budget that year and 19.5% of the amount spent on malignant skin lesions. Hospital plastic surgeons spent the most resources on benign skin lesions in the Cancer Pathway (see Table 4).

TABLE 4.

Economic estimate of the Euros spent in 2021 on benign skin lesions in Cancer Pathway.

Medical specialty	Healthcare setting	Number of biopsies	Average sampling cost per biopsy (range)	Average pathology cost per biopsy (range)	Average total cost per biopsy (range)	Total sampling cost, €1000	Total pathology cost, €1000 (95% UI)	Total cost, €1000 (95% UI)
Primary care physician	primary care	182	81 (53–86)	97 (7–1614)	178 (60–1700)	15	18 (10–37)	32 (24–52)
plastic surgeon	Private practice	483	392 (104–415)	114 (7–1614)	506 (111–2029)	189	55 (51–64)	244 (240–253)
Plastic surgeon	Hospital	567	674 (661–681)	149 (7–1614)	823 (668–2295)	382	85 (52–164)	467 (434–546)
Dermatologist	Hospital	5689	734 (661–752)	168 (7–1614)	902 (668–2366)	4170	958 (720–1532)	5128 (4890‐5701)
Dermatologist	Private practice	2162	198 (123–213)	118 (7–1614)	316 (130–1827)	428	255 (167–468)	683 (595–896)
All	All	9083				5184	1371 (1000–2267)	6555 (6185–7450)

Note: ‘Average total sampling cost’ includes expenses to perform the biopsy, informing the patient of the benign result by letter and suture removal.

Abbreviation: UI, uncertainty interval.

Results of a sensitivity analysis is shown in Figure 5, with details in Table S2.

FIGURE 5.

Sensitivity analysis of the current cost of benign skin specimens sent for histopathological evaluation using Cancer Pathway and four alternative ways of handling these skin lesions, conservatively assuming 25% of lesions in all scenarios with sequential follow‐up were excised. The median cost is indicated by a point, the first and third interquartiles are indicated by error bars and the width of the violin body shows data density. Current costs: The estimated cost of current practice. Scenario 1: The lesion is not immediately excised, but the patient is seen for sequential follow‐up consultations at the place where the biopsy was to be performed to exclude malignancy. Scenario 2: The patient was never referred but had the lesion excised by the Primary Care Physician on the initial visit. Scenario 3: The patient is referred to a dermatologist in private practice and seen initially for one follow‐up consultation using digital sequential dermoscopy. After this, 25% are referred to a private plastic surgeon for excision. Of these, 10% are subsequently referred to a hospital plastic surgery department due to special needs (e.g. large facial lesions) or co‐morbidity. Scenario 4: The patient is followed using two digital sequential teledermoscopic follow‐ups via the primary care physician, after which 25% are referred to a private plastic surgeon for excision. Of these, 10% are subsequently referred to a hospital plastic surgery department due to special needs (e.g. large facial lesions) or co‐morbidity.

DISCUSSION

This study shows that a healthcare system that employs layered filtering functions before biopsy of skin lesions can achieve some of the lowest BMR and NNB reported in the world, ⁷ with an overall BMR as low as 1.5:1, an NNB of 2.8 and an underdiagnosis rate of 1 melanoma in 161 biopsies. However, this comes with the price that most excisions of benign skin lesions are performed in hospitals, which is the most expensive.

The 2007 implementation of the Danish national fast‐track referral system initiative aimed to reduce the time to diagnosis and has, since 2008, included a fast‐track Cancer Pathway for patients suspected of having melanoma with two main entries: For lesions where melanoma cannot be ruled out, the PCP refers the patient to ‘filtering function’ managed by the private dermatologists or plastic surgeons. For lesions strongly suspected of melanoma, the PCP or dermatologist refers the patient directly to a plastic surgery department—before or after a biopsy, which is also expedited under the Cancer Pathway. ³⁵ , ³⁶ , ³⁷ , ³⁸ Up until 2020, excisional biopsy of strongly suspected melanoma was recommended by the Danish Medical Authority to be performed only in plastic surgery departments, which affects the reported BMR and NNB. ³⁴ , ³⁵ Comparing the reported ratios of the included medical settings should take this into account, as most obvious melanomas were referred rather than excised by PCPs and dermatologists in Denmark. This also makes comparison with international literature difficult, as in many countries, dermatologists perform more surgeries themselves, which, of course, will affect the patient flow.

Interestingly, our analysis reveals that the clinicians use the Cancer Pathway in broader terms with a proportional steady state of more than 40% of the malignant tumours in this pathway being BCC, PCC and other malignancies (Figure 3; Table S3). An explanation could be the clinical need for a diagnosis within a definite time frame for optimal treatment of aggressive carcinomas. Similarly, the high rate of melanomas in Routine Pathway biopsies from the hospital plastic surgeons could originate from the practical workflow structures and rapid pathology reports despite pathway choice at the hospitals.

The study includes BMR and NNB analyses for lesions in the Cancer Pathway, where we can confidently assume that the clinicians suspected malignancy or melanoma. This provides a unique insight into biopsy practices across healthcare settings but likely do not reflect diagnostic competence and should not be directly compared to findings in other studies, as these figures are heavily influenced by the organizational patient management guidelines. Lesions clinically suspected of melanoma are easily referred to the next healthcare setting, which was done more often than lesions were excised by PCPs (Tables 1 and 2), in accordance with the guidelines at the time. ³⁴ , ³⁵ This use of referral to the ‘filtering function’ is likely also an explanation for the relatively low NNB in Cancer Pathway of 9.2 found among the included PCPs compared to 22.6 and 14.6 reported in meta‐analyses. ⁷ , ³⁹

A recent US study found a NNB of PCPs of 34.3 and 11.3, before and after an educational intervention and access to teledermatology, respectively. ⁴⁰ This study included only nevi and melanoma. Our more comprehensive array of potential ‘melanoma mimics’, while still achieving a lower NNB, underscores the effectiveness of the healthcare system organization and ‘filtering function’ in accurately filtering melanoma cases amidst a diverse array of skin lesions. The effect of education and access to teledermatology in our PCP population will be the subject of future research.

From an economic perspective, the established patient flow might be an unfortunate practice as the price of a skin biopsy is almost double when performed in the hospital setting compared to the private practice setting. This is reflected in our sensitivity analysis, with all scenarios of managing benign skin lesions being cheaper than the current practice. This is partly because excisions are moved from the expensive hospitals to the private practice setting and partly due to the assumption that only 25% of the benign lesions would be excised if digital sequential follow‐up were used but does not consider the potential delay in melanoma diagnostics from sequential dermoscopic follow‐up. Excision of all benign skin lesions by the PCPs appears cheapest for society. However, given the relatively low diagnostic accuracy of PCPs, ⁴ , ⁴¹ , ⁴² , ⁴³ this would likely also increase the number of biopsies and hereby raise the costs, which is not reflected in the sensitivity analysis.

Limitations

Misclassification of subgroups of specimens, revisions and risk of double counting lesions subjected to multiple biopsies could be present; however, considering the extensive scope of the dataset, such instances are unlikely to impact the study's overall results substantially. Another limitation is the topographical near‐match algorithm for matching histological revisions with the initial biopsy, which has yet to be officially validated. However, the original algorithm has been thoroughly tested and set into production for quality data extraction of the Danish Melanoma Database. As histopathological revision was only performed on a fraction of the included cases, this also likely did not skew the results.

The simple cost analysis was performed without including the associated opportunity costs for patients and healthcare systems, which affects comprehensiveness. Several assumptions underlie the economic calculations. However, we believe them to be conservative. With the use of accurate, deterministic, and national reimbursement rates for both the simple cost analysis and economic sensitivity analyses, we believe our findings realistically reflect the relative cost of current practice and the alternative scenarios.

Among the study's strengths are the comprehensive data material, which includes every biopsy of interest submitted to the Danish Pathology Data Bank for 15 years from the entire country, the variety of diagnoses included as benign melanoma mimics and the transparent and conservative simple cost analyses.

Future perspectives

To enhance the already favourable BMR and NNB, efforts should focus on reducing benign unnecessary biopsies through more effective triage or education. Training PCPs in dermoscopy and providing quick access to expert second opinions can lower their NNB⁴⁰. However, the main economic advantage lies in impacting the NNB among hospital dermatologists and plastic surgeons, who, despite their low NNB, contribute significantly to the economic burden by handling the majority of skin lesions in the Cancer Pathway. Additionally, transferring sequential follow‐up and biopsy procedures to private practices rather than hospitals might yield substantial economic benefits and lower the rate of benign excisions. This shift would likely be driven by the need to save money and resources in the hospitals. As physical consultations with a consultant dermatologist are scarce, teledermoscopic evaluation seems the most feasible option of the four scenarios. How the implementation of such a scenario affects the management of malignant skin lesions and the time to diagnosis has yet to be determined and implementation should not be done without adequate training of PCPs and easy access to competent tele‐triage options, which are essential to maintaining acceptable levels of BMR and NNB.

CONCLUSION

This study shows that a healthcare system that employs easy and swift access to filtering functions before biopsy of skin lesions can achieve a BMR as low as 1.5:1, NNB of 2.8, and an underdiagnosis rate of 1 in 161, but it carries a heavy toll on the public healthcare system of more than €6.6 million spent on benign skin biopsies per year. Additionally, it shows NNB is heavily influenced by clinical suspicion, which should be included in future studies that report diagnostic accuracy using this measure.

AUTHOR CONTRIBUTIONS

All authors participated in the conceptualisation, drafting and revising the manuscript and approved the final version of the article.

FUNDING INFORMATION

The study was funded by the Danish Cancer Society, the Research Fund at Herlev and Gentofte Hospital and the Innovation Fund Denmark.

CONFLICT OF INTEREST STATEMENT

LRH is the chairman of the Danish Melanoma Group and Database. NKT is CEO and founder of MelaTech ApS, a company that provides an eConsultation platform for dermatology. No other authors have conflicts of interest to disclose.

ETHICAL APPROVAL

This study was permitted by the Regional Ethics Committee (R‐22040549) and Legal Department of Scientific Research (2022‐495) at the Capital Region of Denmark.

Supporting information

Figure S1.

Figure S2.

Table S1.

Table S2.

Table S3.

Nervil GG, Vestergaard T, Klausen S, Tolsgaard MG, Ternov NK, Hölmich LR. Impact of skin biopsy practices: A comprehensive nationwide study on skin cancer and melanoma biopsies. J Eur Acad Dermatol Venereol. 2025;39:1267–1277. 10.1111/jdv.20371

DATA AVAILABILITY STATEMENT

Data are patient‐sensitive and cannot be ethically shared. Data processing codes are available from the corresponding author upon reasonable request.