Institutional Review Board approval was obtained for this retrospective study, and informed consent was waived due to the retrospective nature of this study.

Twelve patients were treated with HIFU from October 2008 to May 2010. Table 1 summarizes the patients' data. The 12 patients were determined to be eligible for HIFU treatment by using the following criteria: they had a diagnosis of pancreatic cancer as confirmed by the pathologic findings or by the clinical and imaging findings typical of pancreatic cancer, the tumor was unresectable based on the guidelines of the 6th edition of the American Joint Committee on Cancer (AJCC), the lesion was mainly located in the pancreatic head or body and they had a Karnofsky performance status scale rating of at least 70%.

The following inclusion criteria were used to examine the effect of CCHT: 1) the patients underwent at least three sessions of concurrent treatments, which were defined as HIFU within 24 hours of a gemcitabine injection, 2) initiation of the first CCHT was done within three months of the diagnosis to meet the requirements that concurrent treatment should be the main treatment and 3) there was no history of radiation or cyberknife treatment before and after CCHT. Three of the 12 patients satisfied these criteria and they were categorized into the "CCHT group." The other nine patients were categorized into the "non-CCHT group."

A FEP-BY™ HIFU unit (Yuande Biomedical Engineering Limited Corporation, Beijing, China) was used throughout this study (14). The HIFU treatment was performed using an upper HIFU transducer enveloped in a degassed water bladder with the patient lying supine on the treatment table. The targeted pancreatic tumor was identified using a B-mode ultrasound imaging transducer (GE Logiq 5, Seongnam, Republic of Korea) before the HIFU treatment. The HIFU beam was insonated into the body and moved automatically from spot to spot in an overlapping manner to treat a volume of tissue. Local or general anesthesia was not needed in any of the patients. The patients were fasted for nine hours before the HIFU treatment. The treatment parameters were input target energy: 500-1000 J/spot, input acoustic intensity (spatial average-temporal average intensity, I_sata): 1-2 kW/cm², pulses/spot: 50-70, transit time of a unit pulse (t1): 150 ms and intermission time between pulses (t2): 150 ms, intermission time between spots: 5 sec (duty cycle: 50%, pulse repetition frequency: 3.3 Hz, sonication duration: 15-21 sec). The acoustic power was adjusted with respect to the tumor depth and the subcutaneous tissue thickness, as measured by ultrasound imaging, to achieve the required in situ energy dose (15). To avoid skin burns, a cold ultrasound coupling gel was applied frequently between the water bladder and skin during the procedure.

Gemcitabine (2,2-difluorodeoxycytidine, Gemzar; Eli Lilly & Co., New York, NY) 1000 mg/m² was administered as a 30 minute intravenous infusion weekly for three weeks followed by a one week rest period. Chemotherapy was omitted according to the physician's decision if a significant hematologic or nonhematologic toxicity had not recovered on the day of therapy. Within 24 hours of infusing gemcitabine, the HIFU treatment was conducted using the protocol described above. HIFU was generally cancelled when chemotherapy was cancelled due to an adverse reaction. However, the HIFU treatment alone was occasionally performed in cases with a long absence of treatment. Gemcitabine administration was continued until disease progression occurred.

Regarding the three patients in the CCHT group, the following were analyzed: the number of CCHT sessions, the number of HIFU alone treatments, the mean target energy per spot (J/spot), the mean input acoustic intensity (W/cm²), the mean treatment time from the first to last sonication, the serial changes in the level of the tumor marker CA 19-9, the serial changes of the tumor size by CT, the PET-CT findings if available, the OS from the time of diagnosis, the time to tumor progression (TTP), the presence of complications (redness, skin burn, treatment-related pain, pancreatitis, gastrointestinal injury and others) and the current performance status as determined by the Karnofsky scoring system.

The OS was calculated from the date of diagnosis to the date of death from any cause or the last documented follow-up. The TTP was calculated from the start of treatment to the date of the first documented progression or the last follow-up.

Regarding the nine patients in the non-CCHT group, we determined the main reason why CCHT was not performed more than twice. The OS from the time of diagnosis, the TTP and the presence or absence of complications during the HIFU treatment were also analyzed.

Tables 1 and 2 summarize the survival data, the treatment protocol and the complications of the CCHT group. Figures 1-3 show the serial changes in the CA 19-9 levels and the CT-determined tumor sizes.

Patient 1 (Fig. 1) (Tables 1, 2) had a 2.7 cm soft tissue lesion encasing the superior mesenteric artery (SMA) with an indistinct low attenuation area in the uncinate process of the pancreas, as seen on CT (Fig. 1A). The laparoscopic biopsy revealed atypical ductal cells and this suggested pancreatic carcinoma. PET-CT showed increased activity around the SMA. After gemcitabine was first administered in January 2009, CCHT was performed since January 2009 according to the gemcitabine infusion schedule. At the end of the first six sessions of CCHT, the CA 19-9 level had decreased from 50.8 U/mL to 21 U/mL. After six sessions of CCHT, three consecutive weekly CCHTs were performed with a 1-3 month intermission. The tumor size decreased and it remained stable in size for approximately one year since the first CCHT (Fig. 1B). However, subsequently, the CA 19-9 level began to be reelevated to more than 100 U/mL and the follow-up PET-CT revealed a small region of hot activity at the left paraaortic area that had not been treated with CCHT (Fig. 1D). Afterwards, the paraaortic lesion was treated with three consecutive sessions of CCHT. The follow-up CT (March 2010) showed that the attenuation of the left paraaortic area was reduced compared to the previous CT. On March 18, 2010, a new chemotherapeutic regimen (TS-1 and CDDP) was attempted due to a continuous increase in the CA 19-9 level. However, it was cancelled after the first treatment session due to severe adverse effects. Four consecutive sessions of CCHT with gemcitabine were performed from April 7, 2010 to April 28, 2010. After this treatment, the CA 19-9 level was decreased temporally (1341 U/mL → 1105 U/mL). Currently, the patient is refusing further chemotherapy. In September 2010, at the time of preparing this manuscript, he was still alive in a good physical condition (Karnofsky performance status scale: 80-90%), and there are no sign of distant metastases.

Patient 2 (Fig. 2) (Tables 1, 2) was diagnosed with a 3.3 cm unresectable pancreatic head cancer invading the peripancreatic tissue and the distal common bile duct with enlarged mesenteric lymph nodes according to the CT taken in July 2008. Three days later, drainage of the common bile duct was achieved via endoscopic retrograde cholangiopancreatography (ERCP) through the insertion of an uncovered metallic stent. At that time, the tumor was considered to be surgically resectable. However, the follow-up CT taken in September 2008 revealed an interval increase in the size of the main pancreatic mass (3.3 cm → 4.3 cm) and the mesenteric lymph nodes, and the new development of a small liver metastasis in segment 5 of the liver (Fig. 2A). Therefore, surgery was cancelled and a gemcitabine-capecitabine (Xeloda; F. Hoffmann-La Roche AG, Basel, Switzerland) combination was administered. The capecitabine was administered orally at a dose of 830 mg/m² twice a day with gemcitabine being given once a week for the first three weeks. However, the capecitabine was discontinued due to hand-foot syndrome. CCHT was performed from November 2008. After the first three CCHT sessions, the CA 19-9 level decreased to 40.3 U/mL and the tumor size according to CT decreased to 2.8 cm in diameter (Fig. 2B). The liver nodule observed in segment 5 had completely disappeared according to the CT taken in May 2009. However, the CA 19-9 level was elevated to more than 100 U/mL in June 2009 and the CT in August 2009 revealed a slight interval increase in tumor size (2.5 cm → 2.9 cm in diameter). Surgical resection was considered given the lack of vital vessel involvement by the pancreatic tumor and his excellent physical condition (Karnofsky performance status scale: 100%). The liver lesion in segment 5 was not observed before surgery. After some delay of surgery due to cholangitis and a myocardial ischemic attack, the surgical resection (pylorus preserving pancreaticoduodenectomy) was performed in November 2009 (Fig. 2C). The pathological diagnosis was poorly differentiated adenocarcinoma, and the surgical staging was T3N1. No peritoneal seeding was identified. However, on a postoperative follow-up CT taken in November 2009, a single 1 cm nodule reappeared in segment 5 of the liver and it began to increase slowly in size, confirming that it was a liver metastasis (Fig. 2D). In addition, the portocaval lymph node had also increased slowly in size. A new chemotherapeutic regimen (TS-1 and CDDP) was started from April 2010. Thus far, the patient's Karnofsky performance statue scale is 90-100%, and no signs of another distant metastasis have been observed.

Patient 3 (Fig. 3) (Tables 1, 2) was diagnosed with a 4.6 cm infiltrative unresectable cancer of the pancreatic head with invasion of the distal common bile duct and encasement of both the celiac axis and the main portal vein in August 2008 (Fig. 3A). From the time of the diagnosis, there was a small quantity of ascites, which might represent peritoneal seeding (Fig. 3B). Therefore, the initial TNM staging was T4N1Mx (stage III or IV). In August 2008, ERCP was performed with the insertion of an uncovered metallic stent for drainage of the common bile duct. From August 2008, six sessions of a weekly gemcitabine infusion (2 cycles) were performed with capecitabine, which was administered orally at a dose of 830 mg/m² twice daily for three weeks followed by a one week rest period. However, the CA 19-9 level increased (140 U/mL → 264 U/mL) and no definite change in tumor size occurred according to CT. CCHT was started in October 2008. The tumor was partially treated during every CCHT session because the upper half of the tumor around the celiac axis was always hidden by gastric gas on ultrasound. After the first three consecutive sessions of CCHT, the CA 19-9 level decreased (264 U/mL → 177 U/mL) as did the tumor size from 4.6 cm to 3.7 cm in diameter (Fig. 3B). The ascites also disappeared (Fig. 3C). In May 2009, when approximately two months had passed after his last CCHT, the CA 19-9 level was found to have increased abruptly to 1810 U/mL even though his physical performance status was excellent (Karnofsky score, 90%) and the CT did not reveal any significant interval change in tumor size at that time. In June 2009, his chemotherapeutic regimen was changed to FOLFOX (folinic acid [FOL] fluorouracil [F] and Oxalipatin [OX]), which unfortunately caused severe leucopenia and subsequent biliary sepsis (a known severe adverse reaction of the drug). Sadly, in July 2009, he succumbed to the biliary sepsis.

Of the 32 CCHTs and six HIFU alone treatments administered in the CCHT group, a superficial grade 2 skin burn occurred in patient 3, and a grade 1 skin burn was noted in patients 1, 2 and 3. The grade 2 skin burn measured approximately 4 × 5 cm in size and it required silver sulfadiazine ointment and a sterile gauze bandage for two weeks. It resolved completely within three weeks. The two grade 1 skin burns recovered without treatment within one week. No major complications such as gastrointestinal perforation or pancreatitis were encountered in the CCHT group. Mild abdominal pain was reported during treatment (during 2 sessions in patient 1, during 6 sessions in patient 2 and during 2 sessions in patient 3). Moderate abdominal pain was noted during one session in patient 1, during two sessions in patient 2 and during one session in patient 3. However, the pain was controlled with Fentanyl, and it disappeared immediately after completing the HIFU treatment.

The patients in this group did not undergo CCHT more than twice for the following reasons: intolerance of pain during treatment (n = 4), a poor physical condition due to disease progression (n = 3), palliative use of HIFU for pain relief (n = 1) and pancreatitis (n = 1). The patient who developed pancreatitis after the HIFU treatment had a small cystic portion close to the pancreatic cancer before treatment, which was probably a pseudocyst. This cystic lesion was included in the treatment field. Two weeks after the 3rd HIFU treatment, a large pseudocyst surrounded by inflammatory changes occurred in the mesentery anterior to the pancreas. This complication might have been caused by the delayed perforation of the cyst near the pancreatic tumor due to damage of the cystic wall by HIFU. All four patients in whom the HIFU treatment had been terminated due to severe pain during treatment were treated with more than 900 J/spot. The use of a high HIFU energy (> 900 J/spot) was favored for approximately two months after initiating HIFU treatment at our center, yet it frequently caused severe abdominal pain. Accordingly, the energy delivered was lowered to < 800 J/spot, and patients have since felt more comfortable during treatment. The patient who underwent HIFU for pain palliation did experience improvement of their pain (numeric pain scale 7 → 3).

The median OS and TTP of this group were 10.3 months (95% confidence interval [CI]: 1.4-19.3) and 4.4 months (95% CI: 0.0-9.0), respectively. A grade 1 skin burn occurred in only one of the nine patients, and this resolved within one week without any special treatment. Subcutaneous sclerosis caused by thermal injury to the subcutaneous fat of the anterior abdominal wall occurred in two patients, but this occurred after they had been administered 1000 J/spot. The sclerosis took four weeks to resolve without treatment in one patient. However, in the other, it was not completely resolved four months after completing the HIFU treatment.