We present an imaging methodology to detect the presence of an anechoic lesion in a highly scattering medium consisting of spherical air scatterers. Classical imaging methods have failed to image such media. In the method presented here, the incoherent backscattered intensity is extracted and the linear growth of the diffusive halo is tracked. Sudden changes in this growth indicates the presence of a target. This algorithm combined with the algorithm of Winslow et al. on contour detection enables us to predict the presence and the location of such a lesion. This method has been developed to detect the presence of pulmonary nodules and ground glass opacities in the lung parenchyma. Using a 128-element linear array transducer operating at 5 MHz, experimental results were obtained from a sponge phantom with an air volume fraction of 50%, in which gelatin nodules of diameter 5 mm and 8 mm were implanted at depths of 15 mm and 20 mm, respectively. The diameter of the nodules could be predicted within an error margin of ±25%. FDTD simulations were also carried out with nodule sizes of 6mm and 11 mm, in media with 20% and 50% air volume fractions. The coordinates of the center of the lesion could be predicted within a tolerance of ±7%.