Microchannel flow boiling is an attractive thermal management strategy for ever-growing volumetric heat dissipation demands associated with electronic systems. Due to difficulties related to measurement at the microscales, the majority of researchers have chosen to study relatively simple situations with uniform heat flux. However, many applications involve local hotspots which give rise to highly non-uniform heat flux and temperature gradients due to heat spreading. This necessitates the consideration of conjugate heat transfer for accurate analysis. The current work is aimed at investigating conjugate heat transfer in a two-phase microchannel array. Experimental data was collected on R134a flow boiling heat transfer for very small hydraulic diameter (<100 mu m) silicon microchannels with very high heat fluxes (>1 kW cm(-2)) applied via platinum strip heaters with a footprint size of 1 cm x 1 mm. The collected experimental data was then combined with detailed computational modeling utilizing finite element modeling with COMSOL Multiphysics and MATLAB to examine the applicability of five published heat transfer correlations for use in determining local heat transfer coefficients. A two-phase correlation from Agostini and Bontemps, developed for markedly different test parameters, provided the best computational agreement with an RMS temperature difference from experiment of 3.3 degrees C and a predicted peak perimeter-averaged heat transfer coefficient of 116 MW m(-2) K-1. Modeling confirms the presence of highly non-uniform local heat flux and correspondingly non-uniform local heat transfer coefficient. The results of this study make clear the need for better micro-scale two-phase correlations developed to predict local heat transfer coefficients at these small scales and high local heat fluxes. (C) 2018 Elsevier Ltd. All rights reserved.