Utility of the Singapore recurrence nomogram in a US cohort with a higher proportion of borderline and malignant phyllodes tumours.

Aims: For phyllodes tumours (PT), local and distant recurrence rates increase with higher grades and are difficult to predict. The Singapore nomogram has been used to predict recurrence events for PT. We aimed to test this nomogram for accuracy in a US cohort and to compare with a histological score.

Methods: Patients with PT were selected from a prospective institutional database. Histological parameters and margin status were used to estimate the nomogram score and the histological score, as previously defined. Multivariable analyses were used to estimate the association of recurrence-free survival (RFS) with individual factors, nomogram score and histological score. Harrel's C-index was estimated.

Results: Of 81 PT cases, 25.9% were benign, 40.7% borderline and 33.3% malignant. Recurrences occurred in 33.3% (n=27). The adjusted RFS analysis including the four factors used in the Singapore nomogram performed well (C-index of 0.78). However, despite a higher nomogram score being associated with increased risk of recurrence (HR 1.03, 1.01-1.05, p=0.007), the individual numeric scale defined in the nomogram only moderately fit our data (C-index of 0.66). Patients with higher histological scores also had an increased risk of recurrence (HR 1.25, 1.07-1.47, p=0.005; C-index of 0.70).

Conclusion: Histological score more accurately predicted PT recurrence in our cohort, which includes a higher proportion of higher-grade PT. Refining the nomogram to include factors specific to malignant PT and factors with more variance, as well as refining the assigned weights, may result in improved performance. This study identifies an opportunity for international collaboration to refine the predictive model.