le 1 mars 2012
There is a new world intellectual champion in the field of deformable solid registration. Our paper (1) solves an intractable computing problem when the minimum energy does not correspond to the solution. When 100 parallel computers are linked, it will only get to the wrong solutions 100x faster - truly a fastlane to failure (sorry NSF I borrowed your idea and post my challenge on your walls). I am inviting Northern Europe to challenge my intellectual global dominance, solving an intractable computing problem, no future by-passes (1). An active global dominance is like your neighborhood pizza guy stick his recipe on the pizza guy down the street and screaming I am better than you and going to tell all my customers that. Can you imaging what the other pizza guy will do?
This is the daily email sent out counting the days for a month:
Time = Day 30 of 30 (Feb 23, 2012), for an active global challenge to be #1.
Hi, how are you?
Do you have a faster or know of a more accurate program to solve the problem of large deformations in registering multi-modal images. This is until now an intractable computational problem. Our solution is being integrated into prostate cancer radiation treatment to spare potency. Our solution will benefit the fields of medicine, computing and engineering, so I'm opening this challenge to the colleagues in computer science, applied mathematics and engineering.
My claim = # 1 in solving large deformations in image registration.
I am gathering the answers to this global challenge from more than a few dozen universities in Asia, Europe, North and South Americas, Australia, Middle East, and the Nordic region. I'll start to write it up for publication in the end of Feb, 2012.
I welcome anyone to meet this challenge. This rigorous established the singularity of our algorithm (1). It is anticipated that our work will survive this world-wide challenge.
To demonstrate the big gap between what is commercially available and what is needed in the clinic. Using reference 2 as an example, the MRI information was transferred to the CT scan using fusion. Fair to say, the study was performed using the state of the art fusion algorithms available. Rightly so, the IGRT was done with or with image registration. However, without image registration or simply use fusion (that implies rigid body translations and rotations) are simply subpar and should be wiped out. Your algorithm is by-passable, more computers in parallel will likely bring you down. Please challenge my assertion here. I love to be wrong on this. Anyhow, yours (3) is by-passable, right? Since we have solved deformable solid registration (1) with error estimation. The accuracy of the fusion is judged by eye-balling only. There is no quantification of the additional planning margin required to account for the errors in registration. Our algorithms have extensions in the arena of engineering and applied mathematics. Our implemented algorithms are a part of our Interactive Specialized Prostate Image Registration Ensemble (InSPIRE) available at request, and it is ready for clinical implementation, although it is patent pending. I am wiping clean the treatment planning and multimodal deformable registration. It would get 5 star assuming I were wrong. Publisher, Drs. Cox and Ang, and ASTRO, please do correct me, I would love to hear I were wrong !
Rex Cheung, MD, Ph.D. P&S
Associate Professor of Radiology,
Texas A and M.
Bryn Mawr, Philadelphia, PA.
1. Using manual prostate contours to enhance deformable registration
of endorectal MRI. R. Cheung, K. Krishnan, Computer Methods and Programs in Biomedicine. In press.
2. Variation in external beam treatment plan quality: An inter-institutional study of planners and planning systems. Benjamin E. Nelms, Greg Robinson, Jay Markham, et al., PRO. In press.
3. Critical Appraisal of Acuros XB and Anisotropic Analytic Algorithm Dose Calculation in Advanced Non-Small-Cell Lung Cancer Treatments. Antonella Fogliata, Giorgia Nicolini, Alessandro Clivio, et al. International Journal of Radiation, Oncology, Physics. In press.