Early assesment of radiation response using a novel functional imaging modality-[¹⁸F] Fluorocholine PET (FCH-PET): A pilot study

Abstract

Purpose

[¹⁸] Fluoro-deoxyglucose-positron-emission tomography/computerized tomography (FDG-PET) is commonly used to assess response to patients treated with radiation (RT) or Chemotherapy and RT (CRT). The intent of this pilot study is to explore whether FCH-PET can serve as an early predictive biomarker for early detection of RT/CRT response.

Materials/Methods

Fourteen patients have been accrued and analyzed. The lesions were base of tongue, tonsil, nodes, hypopharynx, maxilla, palate, lung, pancreas, brain, uterine and rectal. There were 16 lesions that were considered target lesion and were followed for correlation between change in FCH-PET SUV_max readings and clinical outcome. Median tumor size was 4.4 cm. Median RT dose was 66Gy. The change in SUV_max (Δ SUV_max) of FCH-PET scans performed before and during RT was correlated with clinical outcome at the last follow-up.

Results

The median FCH-PET SUV_max for the 1^st and 2^nd scans was 6.15 and 4.65 respectively. Fourteen (87.5%) lesions showed a reduction in SUVmax in either a CR/PR (complete/partial response) and 2 lesions showed an increase in SUVmax both of which were determined to be NR. The median percentage change between the 1^st and 2^nd scan was −19.5%. Forty-four percent lesions (7/16) had CR, 44% (7/16) had PR and 12% (2/16) had NR (No response). Median follow-up was 12 months. The results showed a difference between NR and PR, between NR and CR, showed a trend towards significance (p=0.06).

Conclusion

FCH-PET scan demonstrated changes in SUV_max during RT that were predictive of final outcome

Introduction

Definitive RT, chemotherapy (CT), or a combination of the two (CRT) with or without surgical resection are being employed in the management of primary or metastatic malignancies of head and neck (HN), central nervous system (CNS), gynecological (GYN), Gastrointestinal (GI), Genitourinary (GU) and chest, as dictated by the respective stage of the cancer. [¹⁸]Fluoro-deoxyglucose-positron emission tomography/computerized tomography (FDG-PET) is commonly used to assess response to these treatments (1). Several studies have looked at the role of FDG-PET in assessment of RT response and have found a wide range of predictive values (14–91%) (2). In HN cancers, FDG-PET is able to detect residual nodal disease in patients with a partial response to complete response (CR) with a reported sensitivity of 71% to 100%, specificity of 43% to 100%, positive predictive value of 46% to 100%, and negative predictive value of 66% to 100%.(3–9). A time interval of 8–12 weeks after RT or CRT is required for a restaging by FDG-PET to avoid normal tissue edema or inflammation which has been shown to contribute to false-positive and false-negative results (10, 11). Therefore, there is a need for a test that is able to uncover the ultimate RT/CRT response at the early phase of the treatment course. Such a test may render clinical significance by potentially avoiding subjecting the patient to surgical resection and its related toxicities if no tangible therapeutic benefit is being derived.

One promising non-invasive test is [¹⁸F]Fluorocholine PET (FCH-PET). Choline is a precursor of phosphatidylcholine that is a major constituent of membrane lipids. When choline enters tumor cells, it undergoes phosphorylation and after several biosynthetic processes, finally becomes integrated into lecithin, a major component of cell membrane phospholipids. The biosynthesis of cell membranes is rapid in tumor cells because of fast replication of tumor cells and therefore, the choline uptake is higher in these cells. Radiolabeled choline has been shown to be taken up higher in the tumor cells than in fibroblasts and has been shown to correlate with the proliferative activity of tumor cells (12). FCH-PET has demonstrated its favorable imaging characteristics as a diagnostic/staging tool. For example, in musculoskeletal tumors, a comparison of FDG-PET with [¹¹C]-choline-PET showed that [¹¹C]-choline-PET may equivalent to FDG-PET for differentiating malignant from benign tumors, though the advantages of [¹¹C]-choline-PET were shorter examination time and little retention in the bladder (13) that may be particularly beneficial in gynecological cancers (14).

We hypothesized that FCH-PET tumor uptake at 3–4 weeks into RT/CRT will predict final treatment response in patients. We performed a pilot study is to quantify variance in signal changes between FCH-PET images taken before and during RT/CRT and to determine its correspondence with tumor response at a few months after the completion of therapy.

Materials and Methods

The prospective study was approved by the institutional review board (IRB) and patients consent was obtained prior to enrollment in the study. Patients: From April 2009 till June 2010, 14 patients with malignancies of head and neck, lung, esophagus, rectum and brain were included in the study. Stage I-IV was included (AJCC staging, 6^th edition). Patients were staged prior to enrollment using a PET-CT, CT or an MRI and a biopsy was performed to confirm malignancy in all patients. Treatment: Patients were treated with RT/CRT based on the stage, prior history and overall performance status (Table 2). Study Design: After the initial staging scans for all patients (standard of care), FCH-PET scan was performed prior to beginning RT/CRT. A second FCH-PET scan was performed within 3–4 weeks of starting treatment. After completing treatment, each patient underwent imaging to assess response (CT and/or MRI and/or PET/CT) or surgical resection (eg. rectal cancer) as deemed to be the standard of care for each type and stage of cancer. The variation in the two FCH-PET readings was correlated with the final clinical outcome. Wilcoxon rank-sum tests were employed to compare SUV changes between different response groups. FCH-PET image acquisition: The imaging studies were performed at CBIC using a PET/CT system (Discovery LS; GE Medical Systems, Milwaukee, WI) with an in plane resolution of 4.8mm full width at half maximum. A total of 300–400 mBq of ¹⁸F radiotracer was injected intravenously. A PET imaging study was started at 15 min for FCH and at 1 hour for FDG study. For attenuation correction, non-diagnostic CT was acquired just prior to PET scanning. Following the CT acquisition, a PET scan was acquired in 2-D mode (4–5 min/bed and 2–4 beds as needed). Both attenuation-corrected and non-attenuation-corrected images of FCH and FDG PET data were qualitatively compared for uptake in the malignant and normal tissues by two independent nuclear medicine physicians.

Table 2.

Patient Treatment Characteristics

Size (cm)	4.4 (median)
Definitive RT	12 patients
Palliative RT	2 patients
RT dose	66Gy (median)
Concurrent chemo	11 patients

Results

Patients and tumor characteristics

The patient and tumor characteristics are summarized in Table 1. Fourteen patients have been accrued to date and analyzed. There were 8 males and 6 females. The median age was 65 years (range 45–84 years). There were 16 lesions in 14 patients that were considered target lesion and were followed for correlation between change in FCH-PET SUV_max readings and clinical outcome. Median tumor size was 4.4 cm (range 1.9–13cm). Twelve patients received definitive RT and 2 patients (1 lung, 1 maxilla) received palliative RT. Clinical outcome was measured using clinical exam and/or imaging as well as pathological assessment of resected specimens. Response was defined as either complete response (CR), less than complete response-partial response (PR), and no change or progressive disease was described as no response (NR).

Table 1.

Patient Characteristics

Patient #	sex	age	Site	stage	Histology	Follow-up (months)
1	m	57	BOT	IV	Squamous cell	16
			hypopharynx	IV	Squamous cell
			rt LN	IV	Squamous cell
2	m	84	palate	IV	Squamous cell	1
3	m	72	BOT	III	Squamous cell	14
4	m	71	Lung	IV	Adenocarcinoma	5
5	f	62	lung	IIIB	Adenocarcinoma	13
6	f	70	Lung	IV	Squamous cell	3
7	f	62	esophagus	IV	Adenocarcinoma	13
8	m	65	maxilla	IV	Squamous cell	1
9	m	46	BOT	I	Squamous cell	13
10	m	78	pancreas	III	Adenocarcinoma	4
11	m	71	rectal	IV	Adenocarcinoma	12
12	f	65	brain	IV	Gliosarcoma	12
13	f	45	Lung	IIIB	Adenocarcinoma	9
14	f	67	uterine	II	Adenocarcinoma	6

Change in SUV_max of the two FCH-PET scans

The median FCH-PET SUV_max for the 1^st and 2^nd scans was 6.15 and 4.65 respectively (Figure 1). Fourteen (87.5%) lesions showed a reduction in SUV_max in either a CR/PR (PPV 100%) and 2 lesions showed an increase in SUV_max both of which were determined to be NR. The median percentage change between the 1^st and 2^nd scan was −19.5%. Forty-four percent lesions (7/16) had CR, 44% (7/16) had PR and 12% (2/16) had NR. Median follow-up was 12 months (range 1–16 months). Clinical Response and change in FCH-PET readings are shown in Table 3. Wilcoxon rank-sum tests were employed to compare SUV changes between different response groups. The results showed a difference between NR and PR, and between NR and CR, showed a trend towards significance (p=0.06). However, if the sign of the differences between the two SUV_max values is used as variable, Fisher's exact test shows a statistical significant association with the response (p=0.03, for both comparisons: NR vs PR, CR vs NR) thereby suggesting correlation between change in FCH-PET SUV_max during RT and clinical outcome. While we found these results encouraging, we must stress that only 2 samples were collected that had NR as response. A larger collection of samples is thus necessary to confirm the results of our study.

Figure 1.

FCH-PET scans performed prior to RT and then during RT showing treatment response. 1. Sixty-nine year old female with stage IIIB NSCLC, FCH-PET performed prior to RT, SUV_max 6.0 (A) and 3 weeks later, SUV_max of 5.1 (B). 2. Eighty-four year old with stage IV oropharyngeal cancer with FCH-PET SUV_max of 6.5 before RT (C) and 4.9 (D) after 4 weeks of RT. 3. Fifty-seven year old with hypopharynx cancer, with pre-RT FCH-PET SUV_max of 5.3 (E), and 4.4 after 3 weeks of RT (F). 4. Sixty-one year old female with stage III NCLC, showing a pre-RT FCH-PET SUV_max of 2.5 (G), and 4.3 after 4 weeks of RT (H). 5. Seventy-eight year old with un-resectable pancreatic cancer with pre-RT SUV_max of 6.0 (I), and 2.0 after 4 weeks of RT (J).

Table 3.

Lesion #	Site	FCH-SCAN 1 READING	FCH- SCAN 2 READING	% change	Response at completion of RT	Post-RT scan	Clinical outcome after RT
1	BOT	6.70	5.50	−18.00	CR	FDG-PET	Mass not palpable
2	Hypopharynx	5.30	4.40	−17.00	CR	FDG-PET	No mass seen
3	Rt Neck node	3.70	3.60	−3.00	CR	FDG-PET	No mass seen
4	Palate	6.50	4.90	−25.00	CR	CT	Pt died from complications
5	BOT	5.10	3.80	−26.00	CR	FDG-PET	Mass not palpable
6	Lung	6.30	5.00	−21.00	PR	CT	Hemoptysis stopped
7	Lung	2.50	4.30	72.00	NR	FDG-PET	Stable disease
8	Lung	6.00	5.10	−15.00	PR	FDG-PET	Developed DM
9	esophagus	8.50	6.30	−26.00	PR	CT	Dysphagia significantly improved
10	maxilla	9.90	5.30	−47.00	CR	CT	Developed DM
11	BOT	5.20	4.40	−16.00	CR	FDG-PET	Mass not palpable
12	Pancreas	6.00	2.00	−67.00	PR	CT, CA-19-9	Developed DM, persistent disease
13	Rectal	17.60	1.50	−93.00	PR	surgery	Residual disease in surgical specimen
14	Brain	0.20	0.50	150.00	NR/PD	MRI	Developed local recurrence
15	Lung	9.40	8.20	−13.00	PR	FDG-PET	Developed DM
16	Uterus	9.30	5.60	−40.00	PR	FDG-PET	Local recurrence

In addition, there was minimal non-specific uptake noted on the 2^nd FCH-PET scan thereby minimizing false positive rates as seen on FDG-PET.

In addition to the FCH-PET scans performed at 3–4 weeks after starting RT, clinical exams were also performed as surrogate evaluation to check tumor response. For each patient enrolled in the study, the clinical outcome at the last follow-up (f/u) was as follows: Patient 1: Locally advanced HN cancer, FCH-PET showed a 18%, 17% and 3% reduction in SUV_max for each site respectively at 3 weeks of RT. Clinical exam performed at the time of the 2^nd FCH-PET scan showed a reduction in the size of tumors at each primary location and the nodes. Pre-RT FDG-PET showed an SUV_max of 18.7, 20.9 and 15.4 respectively. These treated areas were non-hypermetabolic after completion of CRT. The patient is disease free in the head and neck area at the time of last f/u but has developed a new primary metastatic lung cancer (adenocarcinoma) that has been treated with chemotherapy. Patient 2: Locally advanced soft palate cancer, 3.8cm tumor with a pre-RT SUV_max was 24.5. FCH-PET showed a 25% reduction at 4 weeks of RT. Clinical exam showed a reduction in the size of tumor at the same time. A clinical exam confirmed complete resolution of all tumors at the primary site and nodes. However, patient died 1 month after treatment due to complications arising from his treatment and multiple co-morbidities. Patient 3: Patient with stage IV non-small cell lung cancer (NSCLC) presenting with hemoptysis. Patient received a course of RT that resulted in complete resolution of hemoptysis. FCH-PET SUV_max showed a 21% reduction and a decrease in the size of the left bronchial mass. Patient developed brain metastasis that was treated with stereotactic radiation. Patient 4: Locally advanced BOT cancer with a pre-CRT FDG-PET SUV_max of 8.9. FCH-PET showed a 26% reduction in SUV_max at 4 weeks of CRT. Clinical exam at 4 weeks revealed a reduction in size of BOT. No metabolic activity on post-treatment FDG-PET. Patient disease free at last follow-up (f/u). Patient 5: Patient with stage IIIB NSCLC, received CRT, showed 72% increase in SUV of FCH-PET at 4 weeks of RT, showed a persistent mass unchanged in size (5.5cm) and FDG-PET showing an SUV_max of 5.5, 3 months after RT, compared to FDG-PET SUV_max of 3.4 and size 2.1cm prior to RT. Patient was clinically diagnosed with persistent disease. Patient has developed renal cell cancer at the last f/u. Patient 6: Patient with stage IIIB NSCLC, pre-CRT FDG-PET SUV_max of 19.8. Patient received CRT and FCH-PET revealed a 15% reduction in SUV_max at 3 weeks of RT. Post-CRT FDG-PET showed a SUV_max of 6.0. However, patient developed metastatic disease 2 months after completion of treatment. Patient expired 4 months later from brain metastasis. Patient 7: Patient with locally advanced and metastatic esophagus cancer, with a pre-RT FDG-PET SUV_max of 12.6. Patient received CRT, FCH-PET showed a 26% reduction at 4 weeks. There was significant improvement in dysphagia. The patient was followed by serial CT scans, not the FDG-PET scans at the discretion of the treating physicians and showed a complete resolution of the esophageal lesion. However, there was progressive systemic disease and patient is currently on chemotherapy. Patient 8: Patient with locally advanced and twice recurrent maxillary cancer, treated with CRT to a clinically visible symptomatic lesion in the right inner orbit. At 3 weeks, there was a reduction in FCH-PET SUV_max of 47%. There was significant reduction in the size of the visible lesion at the same time. Patient had a clinical complete response at the completion of RT. Patient sent to hospice shortly after due to progressive systemic disease. Patient 9: Patient with history of anaplastic lymphoma, developed a stage I BOT lesion with a pre-RT FDG-PET SUV_max of 14.1, received RT, showed a 16% reduction in FCH-PET SUV_max at 4 weeks with a clinical reduction in tumor size on palpation. Post RT FDG-PET SUV_max was 3.4. The lesion was not palpable post-RT. Patient disease free (BOT) at last follow-up. Patient 10: Patient with un-resectable pancreatic cancer, treated with CRT, FCH-PET SUV_max showed a 67 % reduction at 4 weeks of CRT. Serum CA-19–9 was 135U/ml pre-CRT, 60U/ml during CRT and 68U/ml at 3 months after CRT. Patient had a CT response with a reduction in tumor size from 2.5 to 1.4 cm. However, patient was still considered un-resectable and decided to discontinue all therapy at last f/u. Patient 11: Patient with locally advanced rectal cancer, pre-CRT tumor 8 cm in size and FDG-PET with an SUV_max of 10.4. Patient received CRT followed by surgical resection. Surgical pathology showed a 4cm tumor with negative nodes. Patient developed a single lung metastasis that was resected. Patient was disease free at the last follow-up. Patient 12: Patient received craniotomy for gliosarcoma (WHO grade IV), received RT and Temdar concurrently after resection. FCH-PET performed 4 weeks into CRT showed a small area within the resection cavity that had 150% increase in SUV_max. MRI performed 1 month after RT showed an increase in enhancement in the same area and the patient ultimately developed recurrent disease in the area. The recurrence was surgically removed. Surgical pathology revealed recurrent gliosarcoma. Patient disease free at the last follow-up. Patient 13: Patient with stage IIIB NSCLC, FDG-PET pre-CRT showed a 5.6 cm right lower lobe mass with an SUV_max of 37. Patient received CRT and showed a 13% reduction in FCH-PET SUV_max at 3 weeks. FDG-PET performed a month after completion of CRT showed the right lung mass of 4.6 cm and SUV_max of 19. Patient underwent salvage pneumonectomy that showed a 3.5cm tumor and one positive lymph node. Patient is disease free locally at the last follow-up but developed brain metastasis for which the patient received whole brain RT. Patient 14: Patient had refused surgical resection of her uterine adenocarcinoma (presented with vaginal bleeding), FDG-PET showed a 13 cm tumor with an SUV_max of 10.4 pre-RT. Patient treated with RT only and FCH-PET SUV_max showed a 40% reduction at 4 weeks. Patient reported complete resolution of vaginal bleeding and reduction in size of the uterine mass at the completion of RT. However, patient refused any surgical intervention and ultimately had a progressive disease at 3 months after RT-increase in tumor size to 15cm and increase in FDG-PET SUV_max to 13.8. Patient decided to get further treatment at another institution.

Discussion

Our study showed an excellent correlation between FCH-PET scan performed during RT with the final clinical outcome assessed clinically or with imaging.

PET utilizes positron-emitting annihilation event that occurs when electrons and positrons collide and vanish with two opposed photons with a precise energy of 511KeV. This sort of annihilation reaction can be demonstrated in natural radio-isotopes such as oxygen-15, fluorine-18, and carbon-11. The radionuclide that is most frequently used in clinical practice is fluorine-18 (¹⁸F) that has a half life of 110 minutes. The sensitivity of PET scan relates to the fact that glucose analog, fluoro-deoxyglucose (FDG) gets trapped in the cell by phosphorylation. Cancer cells demonstrate significantly increased glucose uptake because of increase expression of glucose transporter enzyme GLUT-1, increased hexokinase activity as well as increased glycolysis. The kinetics of labeled glucose (¹⁸FDG) produces increased signal to noise ratio and the intensity of the signal can be measured using the ‘standard uptake value’ (SUV) that normalizes signal size to infused isotope and patient body mass. This rationale is valid for most regions of the body, with the exception of tissues in which glucose metabolism is normally high; these include the brain, heart muscle, skeletal muscle, brown fat and inflammation. FDG tumor imaging is also poor in the region of the bladder, since FDG is excreted into the urine. FDG uptake is also reduced when the blood sugar level is high, as in poorly controlled diabetes.

FDG-PET had certain limitations in assessing RT response early during RT course. In a study performed in head and neck cancer patients to study the accuracy of PET/CT in assessing response to RT, the accuracy of FDG-PET/CT was 85.7%, compared with 67.9% for CT alone. All false-negative and false-positive FDG-PET/CT results occurred between 4 and 8 weeks after treatment. The authors concluded that the metabolic-anatomic information from co-registered FDG-PET/CT provided the most accurate assessment for treatment response when performed later than 8 weeks after the conclusion of radiation therapy (15). A study performed to prospectively assess the value of sequential FDG-PET scans in predicting the response of locally advanced rectal cancer to neo-adjuvant CRT, a 79.2% specificity, 81.2% sensitivity, 77% positive predictive value, 89% negative predictive value and 80% overall accuracy was obtained. In this study, all patients underwent FDG-PET/CT both before CRT and 5–6 weeks after completing CRT (16). A study in tumor bearing mice to evaluate the potential of D-18F-FMT, 18F-FDG, L-11C-methionine (L-11C-MET), and 3'-deoxy-3'-18F-fluorothymidine (18F-FLT), as PET ligands for monitoring early responses to radiotherapy showed that tumor uptake of D-18F-FMT was decreased on day 1 after irradiation at 6, 20, or 60 Gy, and the decrease persisted until day 7. Tumor uptake of 18F-FDG was elevated on days 1 and 3 after irradiation at 2, 6, or 20 Gy, followed by a decrease in uptake on day 7 in mice irradiated at 20 or 60 Gy. Decreased tumor uptake of L-11C-MET was observed only on day 3 after the irradiation. Only for D-18F-FMT were significant positive correlations found between ligand uptakes at all the time points examined and tumor volume on day 14 after various doses of irradiation (17).

Choline based PET scans is being increasingly utilized for management of cancers (18, 19). Choline PET has also been recently used to assess response to hormone therapy in prostate cancers (20). A study published from Germany showed that [(¹¹)C]choline has the potential for use in the early monitoring of the therapeutic effect of docetaxel in a prostate cancer xenograft animal model. The study also indicated that PET with radioactively labeled choline derivatives might be a useful tool for monitoring responses to taxane-based chemotherapy in patients with advanced prostate cancer (21). Choline PET has also been used to assess response to photodynamic therapy in prostate cancer models (22). Biochemically, [¹¹C]Choline is indistinguishable from the naturally occurring choline. However, the half-life of [¹¹C]Choline is only 20 minutes resulting in the limited use in centers equipped with on-site cyclotron. The need for longer lasting tracers led to the development of ¹⁸F Choline (FCH) based molecules with a half-life of 110 minutes.

Our study was a pilot study to determine the variance in SUV_max of FCH-PET and assessing its correlation with RT/CRT response. However, there was no surrogate FDG-PET scan done at the time of second FCH-PET scan and no FCH-PET scan was performed after completion of RT/CRT (limited funding). However, the patients were evaluated clinically or with serum markers during RT or CRT. Additional scans would certainly have made the study stronger. We are planning additional studies during and after RT/CRT to further evaluate accuracy and specificity of FCH-PET to determine the early RT/CRT response.