To obtain a sufficient visible medical image from a computed tomographic-scan exam, it is necessary to pass through a reconstruction step. Therefore, the quality of the obtained image depends essentially on the e�ciency of the applied reconstruction method. In this work, we propose a new method of reconstruction applied and tested on lung images based on the combination of two well-known basics tomographic reconstruction algorithms: Fil- tered Back Projection (FBP)and Ordered-Subsets Expectation Maximization (OSEM). The combination of these two algorithms is followed by a post-processing stage oering better quality image ensured by Gaussian ltering. The use of these two basics algorithms in this hybridization is justied by an evaluation analysis which ensure the best quality parameters for the relative norm error of the simulated projection (dp),the normalized absolute error (NAE), the normalized cross correlation (NCC), the relative norm error of the reconstructed image (df), the mean square error (MSE), the structural content (SC)and the peak signal to noise ratio (PSNR) of the proposed method compared to other literature algorithms. In addition, and to conrm the signicant dierence of , the Dunnett "test t"' was executed on the means of the quality parameters. In fact, the Dunnett "test t" obtained a p- value (� 0:05) for all the means of the quality parameters which attest the superiority of the proposed method.