Original article
Melanoma staging: Varying precision and terminal digit clustering in Breslow thickness data is evident in a population-based study

https://doi.org/10.1016/j.jaad.2018.03.023Get rights and content

Background

Errors in Breslow thickness reporting can give misclassification of T category, an important classifier in melanoma staging.

Objective

We sought to investigate precision (number of digits) and terminal digit clustering in Breslow thickness and potential consequences for T category.

Methods

All first primary and morphologically verified invasive melanomas in Norway between 2008 and 2015 were included. A smoothing model was fitted to estimate the underlying Breslow thickness distribution without digit clustering.

Results

Thickness was reported for 13,057 (97.5%) patients; the median was 1.0 mm (range, 0.09-85). It was reported as whole numbers (15.6%), to 1 decimal (78.2%) and 2 decimal places (6.2%)—thin tumors with more precision than thick tumors. Terminal digit clustering was found with marked peaks in the observed frequency distribution for terminal digits 0 and 5, and with drops around these peaks. Terminal digit clustering increased proportions of patients classified with T1 and T4 tumors and decreased proportions classified with T2 and T3.

Limitations

Breslow thickness was not reported in 2.5% of cases.

Conclusions

The Norwegian recommendation of measurement to the nearest 0.1 mm was not followed. Terminal digit clustering was marked, with consequences for T category. Pathologists, clinicians, and epidemiologists should know that clustering of thickness data around T category cut points can impact melanoma staging with consequent effect on patient management and prognosis.

Section snippets

Data sources

The Cancer Registry of Norway (CRN) has recorded all cancer diagnoses nationwide since 1953. The Norwegian Malignant Melanoma Registry (NMMR) was established under the CRN in 2008, adding Breslow thickness and other histopathologic and clinical information to each CM case. We included all patients with a morphologically verified first primary invasive CM in Norway between 2008 and 2015 and with Breslow thickness recorded in the NMMR.

Extracting data from cancer-specific registries and working

Results

Between 2008 and 2015, 13,386 Norwegians were diagnosed with a morphologically verified first primary invasive CM. The mean age at diagnosis was 62.8 years of age (range, 2-98 years of age). Breslow thickness was recorded for 13,057 (97.5%) of these patients (6470 men and 6587 women) with a median of 1.0 mm (range, 0.09-85 mm). Thickness was reported to 1 decimal place for 10,211 of the patients (78.2%; range, 0.1-25.5 mm) but also as whole numbers (n = 2032, 15.6%; range, 1-85 mm) and with 2

Discussion

In this national study of Breslow thickness data, thin tumors were reported with more precision than thicker tumors. Reporting of thickness to the nearest 0.1-mm increased by calendar year. Terminal digit clustering was found with marked peaks in the observed frequency distribution for terminal digits 0 and 5, and with drops around these peaks. Smoothing of the observed Breslow thickness distribution demonstrated that terminal digit clustering increased proportions of tumors classified as T1

References (29)

  • L. Ge et al.

    Breslow thickness measurements of melanoma around American Joint Committee of Cancer staging cut-off points: imprecision and terminal digit bias have important implications for staging and patient management

    Ann Surg Oncol

    (2016)
  • S.J. Hayes

    Does terminal digit preference occur in pathology?

    J Clin Pathol

    (2008)
  • A.A. Plumb et al.

    Terminal digit preference biases polyp size measurements at endoscopy, computed tomographic colonography, and histopathology

    Endoscopy

    (2016)
  • A.J. Emmerson et al.

    Rounding of birth weights in a neonatal intensive care unit over 20 years: an analysis of a large cohort study

    BMJ Open

    (2013)
  • Cited by (0)

    Funding sources: None.

    Conflicts of interest: None disclosed.

    Presented as a poster at the World Conference of Melanoma, Brisbane, Australia, October 18-21, 2017.

    Reprints not available from the authors.

    View full text