Distraction osteogenesis is used to treat acquired or congenital limb deformities secondary to tumor, infection, or trauma. To lengthen and/or realign a bone, an external fixator is secured to the segment to be transported via trans-osseous pins that stabilize the bone, followed by an osteotomy through which the bone is distracted. While distraction osteogenesis has yielded positive clinical outcomes for a range of skeletal pathologies, the prolonged process of lengthening and waiting 2x that number of days over which the bone was lengthened for the interposed callus to sufficiently ossify to support body weight is taxing. Complications (infection, contractures, neurovascular injuries, pseudoarthrosis) are correlated with the duration of fixator use. Premature fixator removal results in fracture or recurrent limb deformity. Clinicians make subjective assessments whether the regenerate bone is sufficiently rigid to support weight based on sequential biorthogonal radiographs obtained every 4-6 weeks. Guidelines for determining whether fracture callus is strong enough to support body weight are predicated on the presence of at least three continuous cortices, ≥2mm thick, measured on bi-planar radiographs. However, this calculation is subjective, with <​50% inter-observer agreement. Using this criterion, reported fracture rates after fixator removal range from 3% to 50%. This variation reflects the inability of 2D bi-planar radiographs to objectively portray the 3D spatial distribution of mineralized callus responsible for its structural stability.

Presented in this thesis is an X-ray based machine learning model inspired from previously developed CT model that predicts on the hypothesis that the structural rigidity of a regenerated bone tissue provides of a mechanical assay of the progressive changes in regenerate tissue material properties and anatomical geometry that evolve during distraction osteogenesis and fracture healing. The algorithm accepts biplanar radiographs as inputs, followed by an image processing algorithm to get rid of unwanted noise. Bone Mineral Density (BMD) is predicted from biplanar radiographs, and a Reconstruction Algorithm then predicts 3-Dimensional shape of femur. We have developed an accessible, cost-effective, point of care technology, X-RAY Based Rigidity Analysis Method, which utilizes low-radiation, sequential, biplanar radiographs through fracture callus and regenerate bone to predict its load bearing capacity and failure risk based on calculating the minimal axial, flexural, and torsional rigidities.

In the tranquil embrace of a cozy fall afternoon in 2019, my world was suddenly jolted by the disheartening news of my mother's battle with cancer. The mere utterance of those words shattered the tranquility around me, leaving my heart fragmented into a million pieces. Witnessing her brave journey through treatments, each step laden with both physical and emotional burdens, pierced me to the core.

In the midst of this anguish, a profound determination ignited within me—a resolve born from the depths of love and empathy. It was then, amidst the anguish and uncertainty, that I made a solemn vow to dedicate my life to the pursuit of medical research, particularly in the realm of oncology. My aspiration crystallized:​ to stand as a beacon of hope for those traversing the arduous path of cancer, to alleviate their suffering, and to contribute tirelessly to the advancement of medical science in the quest for more effective treatments and ultimately, a cure.

Thus, as I embark upon this scholarly journey, it is with profound gratitude and unwavering resolve that I humbly present this thesis. It stands not only as a testament to my academic pursuits but also as a tribute to the countless individuals, like my mother, who confront the formidable adversary of cancer with unyielding courage. May the findings within these pages serve as a testament to the enduring spirit of resilience and hope that defines our collective battle against this insidious disease.