Towards multidimensional radiotherapy (MD-CRT): biological imaging and biological conformality

Abstract

Purpose: The goals of this study were to survey and summarize the advances in imaging that have potential applications in radiation oncology, and to explore the concept of integrating physical and biological conformality in multidimensional conformal radiotherapy (MD-CRT).

Methods and Materials: The advances in three-dimensional conformal radiotherapy (3D-CRT) have greatly improved the physical conformality of treatment planning and delivery. The development of intensity-modulated radiotherapy (IMRT) has provided the “dose painting” or “dose sculpting” ability to further customize the delivered dose distribution. The improved capabilities of nuclear magnetic resonance imaging and spectroscopy, and of positron emission tomography, are beginning to provide physiological and functional information about the tumor and its surroundings. In addition, molecular imaging promises to reveal tumor biology at the genotype and phenotype level. These developments converge to provide significant opportunities for enhancing the success of radiotherapy.

Results: The ability of IMRT to deliver nonuniform dose patterns by design brings to fore the question of how to “dose paint” and “dose sculpt”, leading to the suggestion that “biological” images may be of assistance. In contrast to the conventional radiological images that primarily provide anatomical information, biological images reveal metabolic, functional, physiological, genotypic, and phenotypic data. Important for radiotherapy, the new and noninvasive imaging methods may yield three-dimensional radiobiological information. Studies are urgently needed to identify genotypes and phenotypes that affect radiosensitivity, and to devise methods to image them noninvasively. Incremental to the concept of gross, clinical, and planning target volumes (GTV, CTV, and PTV), we propose the concept of “biological target volume” (BTV) and hypothesize that BTV can be derived from biological images and that their use may incrementally improve target delineation and dose delivery. We emphasize, however, that much basic research and clinical studies are needed before this potential can be realized.

Conclusions: Whereas IMRT may have initiated the beginning of the end relative to physical conformality in radiotherapy, biological imaging may launch the beginning of a new era of biological conformality. In combination, these approaches constitute MD-CRT that may further improve the efficacy of cancer radiotherapy in the new millennium.

Introduction

Radiological images have played an important role in medicine since the discovery of X-rays by Roentgen in 1896. In the last two decades, the advent of computed tomography (CT) and magnetic resonance imaging (MRI) has brought a quantum leap to our ability to visualize the human anatomy. Indeed, these advances have directly and significantly improved cancer diagnosis and treatment. In addition, advances in nuclear medicine imaging, with the widespread implementation of single photon emission computed tomography (SPECT) and the emerging widespread availability of positron emission tomography (PET), hold tremendous promise for improving the management of human cancer.

Recent developments in cancer detection, diagnosis, and treatment have suggested new types of images and intensified our need for them. This realization evolved with improvements in imaging and radiotherapy treatment technology, the spectacular advance in our knowledge of cancer at the molecular level, and the cross-fertilization of the multiple disciplines. Whereas up to the present, radiological images are largely anatomical, the new types of images can provide biological and mechanistic data, for example, metabolic information from PET scanning with fluorodeoxyglucose (FDG) radiolabeled with ¹⁸F and functional/metabolic data from nuclear magnetic resonance imaging and spectroscopy (MRI/MRS) studies. Under development are potential methods to characterize the genotype and phenotype of tumors by noninvasive molecular imaging 1, 2, 3, 4.

Given the wide spectrum of information that the “new” imaging techniques can unfold, we suggest the descriptor “biological” for this class of images (in contrast to anatomical). Biological images broadly include those in the metabolic, biochemical, physiological, and functional categories, and they should also encompass molecular, genotypic, and phenotypic images presently under investigation. For radiation therapy, images that give information about factors (e.g., tumor hypoxia, T_pot) that influence radiosensitivity and treatment outcome can be regarded as radiobiological images.

Biological images are needed and useful for many reasons. Images that yield genotype and phenotype information would be helpful for genetic and molecular diagnosis, and for gene therapy. MR functional imaging of the brain may help surgical and other forms of local–regional therapy to avoid critical neurological structures (5). In radiotherapy, the advent of three-dimensional conformal radiotherapy (3D-CRT) and intensity-modulated radiotherapy (IMRT) has escalated the need for images, and in particular biological images, as will be discussed 6, 7.

In the last half of this (soon to become past) century, a number of major technological advances have significantly impacted upon the practice of radiotherapy and incrementally improved its therapeutic efficacy. These include the Co-60 teletherapy unit, medical linear accelerators, treatment simulators, afterloading and remote afterloading techniques, radium and radon substitutes, and computerized treatment planning. In the last decade, patient-specific 3D images are increasingly applied to treatment planning, 3D-CRT planning systems are maturing and becoming widely available, computer-controlled systems and multileaf collimators are emerging as standard features in medical linear accelerators, and the use of electronic portal imaging devices is gradually becoming established. Most recently, the new buzz words are inverse planning and IMRT, advances that have the potential of delivering exquisitely conformal dose distributions to the treatment target relative to the dose-limiting normal structures 6, 7.

At Memorial Sloan-Kettering Cancer Center (MSKCC), inverse planning and IMRT with dynamic multileaf collimators have been clinically implemented since October 1995 6, 7, 8, 9, 10, 11. Initially, the inverse planning algorithm was that of Bortfeld, adapted and incorporated into the Memorial Sloan-Kettering Treatment Planning System 8, 9. More recently, the method of Spirou and Chui has been applied clinically 10, 11. In their approach, the conjugate gradient method is applied to minimize an objective function that is basically the deviation from the desired dose in the various structures of interest (10). Depending on the specific structure, the associated objective function can be cast as a hard constraint, a soft constraint, or a dose–volume constraint. As of the fall of 1999, over 600 patients with cancers of the prostate, head/neck, and breast have been treated with IMRT at MSKCC. Operationally, the modality is well-accepted. In terms of radiation dose, our experience indicates a precision of about 1%. Although the potential clinical benefits associated with the increased dose conformality must be evaluated in outcome studies, preliminary results in terms of reducing treatment-related morbidity and increasing local control appear promising.

Indeed, current technology for delivering conformally shaped external beam radiation therapy may have exceeded our ability to localize tumors and normal tissues by conventional imaging techniques. In particular, IMRT can produce isodose distributions capable of delivering different dose prescriptions to multiple target sites with extremely high dose gradients between tumor and normal tissues. Consider, for example, Fig. 1 which demonstrates an ultraconformal external beam treatment of the prostate. In this example, we assume that biological imaging techniques exist which permit localization of regions within the prostate that contain highest tumor burden (gross target volume [GTV], outlined in orange), and also identification of urethra (blue outline) and prostatic capsule (planning target volume [PTV], green outline). We show in Fig. 1 the exquisite capability, using the existing MSKCC inverse treatment planning system, to sculpt the dose to the desired shape. An analysis of the associated dose–volume histograms (DVHs) demonstrates that the goals of the treatment design are well met: namely, to deliver 99 Gy to the GTV, 91 Gy to the PTV, no more than 86 Gy to urethra, and no more than 76 Gy to no more than 30% of the rectal wall. This treatment plan is currently achievable using a 10-field IMRT treatment technique with 15 MV photons.

The ability of IMRT to “paint” (in 2D) or “sculpt” (in 3D) the dose, and to produce exquisitely conformal dose distributions within the constraints of radiation propagation and scatter, begs the “64 million dollar question” as to how to paint or sculpt. Our hypothesis is that noninvasive biological imaging may provide the pertinent information to guide the painting or sculpting of the optimal dose distribution. There has been much laboratory and clinical research on the so-called predictive assays, i.e., the use of radiobiological characteristics such as proliferative activity (potential doubling time [T_pot]), radiosensitivity (the surviving fraction at a dose of 2 Gy [SF₂]), energy status (relative to sublethal damage repair), pH (a possible surrogate of hypoxia), tumor hypoxia, etc. as prognosticators of radiation treatment outcome (12). Noninvasive biological imaging is an incremental advance in the same direction and may provide “radiobiological phenotypes” (in the nomenclature of the “new” biology) of individual tumors. Important for IMRT, the spatial (geometrical) distribution of the radiobiological phenotypes is also provided by such images, based on which dose distribution may be designed conforming to both the physical (geometrical) and the biological attributes.

Given the above, the purpose of this report is to give an overview of the development of the various forms of biological imaging in the context of the needs of radiotherapy. These advances will be grouped into three categories: MRI/MRS, biological imaging with PET, and molecular imaging/radiobiological phenotyping. (Given that some of these approaches are in early development stage and others in preclinical or clinical testing, it is neither our intent nor possible to give a detailed review and comparison of the relative merits of the different options.) Following these discussions, we introduce the concept of biological target volume (BTV) based on biological imaging, thus integrating physical and biological conformality and leading to multidimensional conformal radiotherapy (MD-CRT).