In modern radiotherapy using ultra-tight treatment margins to spare normal tissues, accurate target delineation is very important. Traditional 2-Dimensional (2D) simulators we used for decades  are now replaced by computed tomography (CT) and magnetic resonance imaging (MRI) . Modern radiotherapy relies on 3D imaging data  that require contouring of the gross target volume (GTV), clinical target volume (CTV) and planning target volume (PTV). GTV includes the tumor imaged and all other pertinent information, CTV includes clinical at risk area of microscopic spread and lymph nodes at risk, internal target volume (ITV) includes CTVs from different respirator phases, and PTV includes the set up errors .This paper is part of a series discussing some of the challenges and solutions of modern radiotherapy planning. For specific cancer sites, this study primarily focused on important aspects related to treatment planning of head and neck, and breast cancers as specific cancer sties .
Observer variability in targeting and contouring
There are well-known inter-observer and intra-observer variations for contouring of target [4-7] and normal tissues [8,9]. For example, in one study, the median PTV and the ratio of the largest to smallest contoured volume were respectively 9.22 cm3 (range, 7.17 – 14.3 cm3) and 1.99 for pituitary adenoma , and 6.86 cm3 (range 6.05 – 14.6 cm3) and 2.41 for meningioma. Some plans used 1-2 mm PTV margin, some used 0 mm margin is this study . When the contours superimposed onto the “ideal” plan, there is an excessive dose of 23.64 Gy (up to 268% of the default plan) in pituitary adenoma and 24.84 Gy (131% of the default plan) in meningioma to the optic nerve
. The optic tract dose was to be kept within 50 Gy (in equivalent dose of 2 Gy fractions). Overall, contouring variability and errors are the most important challenge to modern 3D and 4D radiotherapy treatment planning .
Image fusion and contouring variability because of imaging modalities
Understanding the strengths and limitations of various imaging modalities is important in modern radiotherapy treatment planning. PET/CT has been found to be very useful in contouring of GTVs in treatment planning. When a dedicated PET-CT is used, the target volume could be directly contoured on the PET-CT . Otherwise the PET-CT needs to be fused to the treatment CT, by registering the CT of the PET-CT to the treatment CT in most cases . The information from the registration is used to bring PET of the PET-CT to fuse with the treatment CT . It has been found that using the 40-50% standardized uptake value (SUV) is the best in contouring based on phantom  and 20–40% in some clinical studies . Part of the PET SUV blurring comes from respiratory motion . In one study , 50% PET value and CT lung window and level of 1600 and -300 Hounsfield Units (HU) when the cancer is in the lung correlated best the pathologic size of the cancer. Mediastinal window and level of 600 HU and 40 HU were used when the tumor was close to the mediastinum . The contouring on the fused PET image could facilitate contouring on the CT and MRI that are more commonly used imaging modalities in modern radiotherapy treatment planning [2,13,14]. Cone beam CT used for on-board imaging has poor tissue contrast making it more difficult to contour especially for pelvic tissues [15,16]. MRI (3D and 2D) data can better imaging the pelvic tissues [2,17-19] but would require more complex imaging during radiotherapy.
Consensus and challenges in target and normal tissue contouring
The Radiation Therapy Oncology Group (RTOG) [7,20,21] and other US and international radiotherapy groups [7,22,23] have developed site specific contouring atlases since around 2009, when Intensity Modulated Radiotherapy (IMRT) and 3-Dimensional Radiotherapy (3D-CRT) became the standard of care. Blood vessels were found to be a good surrogate for lymph nodes and CTVs . Contouring in 3D radiotherapy era is very time consuming, for example, the average time in contouring an oropharyngeal cancer case was about two hours . It may be even more time consuming when 4D CT scan data are used . Auto-contouring and semi-automatic programs have recently been developed to save the clinician time in contouring [27-32], including some specialize on using RTOG consensus atlases . However, these auto-contouring programs still need to be validated before clinical use. Other than time consuming, 3D target and normal tissue contouring remain to have many challenges as discussed in this paper.
Head and neck targeting and contouring
The IMRT for head and neck cancer is relative new, only about 10 years . It produced equivalent clinical outcome compared to the large amount of clinical data accumulated in the 2D era , but with less toxicity mostly better salivary function . Usually, elective head and neck radiation treatment is used when the nodal recurrence is about 15-20% [1,25]. The RTOG guidelines for head and neck contouring are limited to N0 disease. In this study , the contouring variability of a node positive (N+) head and neck patient was studied. It was found that, the target to treat in oropharynx has shown significant variability even for published head and neck IMRT experts . Some investigators have chosen to treat ipsilateral neck only for advanced stage III tonsillar cancers, the doses were variable ranging from 66 Gy to 70 Gy in 2 Gy fractions .In this study , 8 out of 20 academic and community centers used one level of CTV dose level, twelve used two CTV dose levels(high-risk and low-risk CTVs). Five of these twelve centers used an expansion of GTV and used the same dose level as used for the GTV . The mean target volume irradiated was 250 cc (range 37– 676 cc) . All centers covered the levels II and III, 95% covered the retropharyngeal lymph nodes and 85% covered the ipsilateral level Ib nodes . Average CTV to PTV expansion was 4.11 mm (0 – 15 mm) . The average time clinicians contouring the target and at risk organs was about two and a half hours . Some clinicians use concurrent chemotherapy , but significant number of centers did not. Thus treatment variability remains a challenge even for a typical oropharyngeal cancer. Most of the information on normal tissue tolerance has been collected with 3D conformal radiotherapy [34-36]. Only, recently head and neck normal tissue tolerance in the IMRT era has been published . More normal tissue tolerance information is needed in the IMRT era.
Breast cancer targeting and contouring
Breast cancer surgery has become more conservative over the past few decades moving away from radical mastectomy . Recently more breast cancer patients undergo immediate breast reconstruction that could present a challenge to post-mastectomy radiotherapy [38,39]. Breast reconstruction using tissue expanders is associated with capsular contraction and other complications after radiotherapy [40,41], autologous more vascularized transplant is more appropriate when radiotherapy is planned . Irradiated flap could atrophize about 21% while 16% in non-irradiated refs over 6-10 months . In a randomized trial RESTORE-2, tissue defect up to 150 ml are eligible for the surgical reconstruction . Recently, DIEP (deep inferior epigastric perforator) flap has been used in additional to TRAM (transverse rectus abdominismyocutaneous flap) [42-44]. In one study, a higher rate of complication was observed in the minority of patients who received 10 Gy scar boost . However, overall post mastectomy radiotherapy (PMRT) to the usual dose of about 50-50.4Gy in 1.8-2 Gy fractions can be safely used after immediate or delayed breast reconstruction [40-45]. For reconstructed breast lumpectomy cavity, and with surgical clips placed, the lumpectomy cavity is contoured as the GTV and CTV is the same as the GTV  as compared with traditional lumpectomy cavity contouring when seroma is contoured as a GTV, and GTV to CTV of about 1 cm is typically used . MRI can better see the lumpectomy cavity because of the fluid intensity on T2 MRI .
In modern era, the side effects of post-operative breast cancer radiotherapy include cardiac toxicity, arm lymphedema, pneumonitis, neuropathy, skin changes . Trastuzumab is associated with cardiac toxicity (1-4%), unlike anthracyclin-related cardiomyopathy, it is not dose dependent, rarely causes death, and reversible when treated or when the drug is discontinued . Trastuzumab is not used concurrently with anthracyclin because of cardiac toxicity . Breast cancer radiotherapy after 1990 using modern techniques showed similar advantages for left sided versus right sided breast cancer patients [1,48]. Pericarditis and related pericardial effusion have decreased from 20% to 2.5% using modern 3D techniques . Local regional radiotherapy is related to 4.1% of radiation pneumonitis, and 0.9% for local radiotherapy , 3.9% when treated with chemotherapy and 1.4% without chemotherapy . Using supraclavicular and axillary fields has a 9-58% of arm lymphedema, the rate is negligible when these fields are not used . Skin thickening and fibrosis occur in 1/3 of patients and in 5% for severe fibrosis . Arm lymphedema occur in 13% of patients after lymph node dissection versus 1-3% using sentinel lymph node dissection . Thus sparing the cardiac and other normal tissues, by following the recommended organ at risk (OAR) guidelines, has become an important area of investigation over the past 2-3 decades and remains to be very important .
For hypofractionated whole breast radiotherapy, 40 Gy in 2.67 Gy fractions is usually used, 9 Gy in 3 fractions is used to boost the tumor bed . In one study , patient is immobilized by a wing board and other personalized immobilization device . CTV includes the whole breast tissue, and is expanded 5 mm to get the PTV . Heart and lungs are contoured as OARs . The median breast volume was 760.64 cc (range 44.77 – 1892.1 cc) . The median boost volume was 143.33 cc (23.07–230.02 cc) . Median time to first skin reaction was 12 days (5 –40 days) . Other dose fractionations have also been used in whole breast hypofractionated radiotherapy and have similar outcome and cosmesis when compared with standard fractionation [49,50]. American Society of Therapeutical Radiation Oncology (ASTRO) recommended patients older than 50 years old, T1-T2N0, without chemotherapy, and dose homogeneity less than <7% are appropriate for hypofractionated whole breast radiotherapy . The use of boost was associated with acute and late skin toxicities .
Over the last couple of decades, much has been learned about targeting the correct treatment volumes, the use of multi-modal image fusion to aid contouring and using advanced simulation (e.g. 4D), and image guided radiotherapy. However, as discussed above, this will be a continuous process along the progress in radiotherapy.
Cranes to Moons: Cranes present good wishes, happiness, and legend has it that each crane lives a thousand years. May each origami crane represent 100 best wishes for each of our patients.